helpfn
bcs
SentimentAnalysisinTwitteFromClassification.pdf
Sentiment Analysis in Twitter: From Classification to Quantification of Sentiments within Tweets
Mondher Bouazizi Graduate School of Science and Technology
Keio University Yokohama, Japan
Email: [email protected]
Tomoaki Ohtsuki Department of Information and Computer Science
Faculty of Science and Technology, Keio University Yokohama, Japan
Email: [email protected]
Abstract—Twitter is attracting significant interests from the research community in the last few years. Sentiment analysis of tweets is among the hottest topics of research nowadays. State of the art approaches of sentiment analysis present many shortcomings when classifying tweets, in particular when the classification goes beyond the binary or ternary classification. Multi-class sentiment analysis has proven to be a very challenging task. This is mainly for the simple reason that a tweet usually does not contain a single sentiment, but many ones. In this paper, we propose a pattern-based approach for sentiment quantification in Twitter. By quantification, we refer to the detection of the existing sentiments within a tweet and the detection of the weight of these sentiments. In a first step, we classify tweets into positive, negative, or neutral. Our approach reaches an accuracy of 81%. We then perform the sentiment quantification on the sentimental tweets (i.e., positive and negative ones) to extract the sentiments within them: we define 5 positive sentiment sub-classes 5 negative ones and detect which exist in each tweet. We define 2 metrics to measure the correctness of sentiment detection, and prove that sentiment quantification can be a more meaningful task than the regular multi-class classification.
I. INTRODUCTION
The term “sentiment analysis” is wide and covers many aspects related to data mining. However, in its most basic meaning, it refers to the identification and aggregation of attitudes, opinions and emotions towards a product, service, issue, event, etc., and the evolution of these over the time. Sentiment analysis, in this context, have been deeply studied and many approaches were proposed in the literature to per- form sentiment analysis, typically to review-related websites.
However, with the rapid growth of social networks, and microblogging website, these websites have presented a richer source of data to study and to dig for many purposes including sentiment analysis. Twitter, as a microblogging website, is one of the biggest web destinations and an open space for people to post and discuss their opinions, thoughts or issues. It has attracted the attention of researchers, in particular for the wide use of hashtags and for the limited length of tweets (i.e., texts posted by Twitter users). The task of sentiment classification in Twitter, mostly binary and ternary classification, has been subject to many studies, and a variety of features were pro- posed to perform the classification.
In this paper, we define a new task related to, yet quite different from, the conventional task of sentiment classification
which is the quantification of sentiments within tweets. We propose an approach that relies on writing patterns, and special sentimental unigrams to perform ternary classification of tweets, and then quantify the existing sentiments within the emotional tweets (i.e., tweets which are classified as “positive” or “negative”). We show how the quantification of sentiments makes more sense than the classification and propose a metric to measure the correctness of quantification. The main contributions present in this paper are as follows:
1) We propose a set of pattern-based features, along with special unigram-base features to perform the ternary classification of tweets.
2) We introduce the task of quantification of sentiments within a tweet, and propose and approach to perform this task.
3) We present an efficient metric to measure the correctness of quantification, and use it to evaluate and show the performances of the proposed approach.
The remainder of this paper is structured as follows. Section II presents our motivations and describes some of the related work. Section III describes in details the proposed method. Section IV illustrates our experiments and results, and Section V concludes this work and proposes possible directions for future work.
II. MOTIVATIONS AND RELATED WORK
A. Motivations
Twitter sentiment analysis has been a challenging task. Ghag et al. [1] cited “Hidden Sentiment Identification”, among others, as one of the most difficult and still open to solve, issues that make Twitter sentiment analysis a quite difficult task. “Hidden Sentiment Identification” refers to the detection of the exact sentiment of the tweet rather than its polarity. This being the subject of [2], we concluded that this remains a challenging task because a tweet usually might contain more than one sentiment.
The object of the current work is to quantify, within each tweet, the different sentiments that are present in it: although tweets are very limited in length (i.e., only 140 characters are allowed), they are mostly emotional and present usually more than one sentiment. A typical example is given in the following two tweets:
978-1-5090-1328-9/16/$31.00 ©2016 IEEE Authorized licensed use limited to: University of the Cumberlands. Downloaded on July 24,2021 at 03:23:21 UTC from IEEE Xplore. Restrictions apply.
∙ “Thank you very much my dear. I like it. This is the best present I’ve ever got ♥”
∙ “Damn it. I hate when this happens. Why does it always crush when I forget to save my work !!”
The first tweet presents sentiments of love, happiness and gratitude. However, the second presents sentiments of anger and hate. This being the case of tweets in general, it might be more interesting to detect these sentiments, rather than just classifying the tweets into positive and negative, or even into different sentiment classes.
As mentioned above, in [2] we proposed an approach that performs multi-class sentiment analysis. The objective of the work is to detect the main sentiment within the tweet among 7 different sentiments. The approach, although it presents good results, highlighted an important challenge: when different sentiments are present within a tweet, it is very hard to perform the classification, even by humans. Therefore, a more interesting task might be the detection of all these sentiments, and the quantification of them within the tweet.
B. Related Work
Twitter has been subject to different studies in the last few years. Most of this work has been focusing on the content of the tweets and how to extract opinions of users and the sentiment polarity expressed in the tweets towards specific topics or objects. Boia et al. [3] and Manuel et al. [4] proposed two approaches that, respectively, rely on emoticons to detect the polarity of tweets and on slang words to assign a sentiment score to online texts. Bouazizi et al. [5] proposed a minimal set of features (i.e., 13 features) that allows to obtain an accuracy equal to 87% for tweets of different topics. Nevertheless, other works have been proposed to use sentiment analysis in Twitter as a tool for predicting the results of elections [6], the current state or the future changes in Oil Busines [7], etc.
A few number of works were conducted on the multi- class classification task such as [8] [9], where an approach that classifies documents into reader-emotion categories is proposed, and [10], where an emoticon recommendation sys- tem that recommends emoticons for posted texts to help to author decide which emoticon to insert is proposed. In [2], we proposed an approach for multi-class sentiment analysis that classifies tweets into 7 different sentiment classes. The approach proposed is scalable and can be run to classify tweets into more sentiments given enough training data are existing. The results obtained are promising, however, they were enough to conclude that a more interesting task might be to detect and quantify all the existing sentiments within the tweet.
III. PROPOSED APPROACH
A. Problem Statement
Given a set of tweets, we aim to classify each one of them depending on their sentiment polarity into positive, negative and neutral. Tweets classified as positive are attributed then sentiment scores representing each the weight of one of the positive sentiments we used, and tweets classified as negative are attributed negative scores representing each the weight of
TABLE I DATA SETS STRUCTURE
Set Positive Negative Neutal Pattern Extraction Set 7528 6435 3815 Training Set 5769 3102 1129 Test Set 4856 2462 682
one of the negative sentiments we used. Therefore, to perform the classification, from each tweet, we extract a set of features, refer to a training set and use machine learning algorithms. We then perform the quantification. Afterwards, we evaluate and analyze the obtained results.
B. Data
From a publicly available dataset1, we collected 40,000 tweets, half of them are classified as positive, and the rest are classified as negative. We asked the annotators to attribute to them any of the following 12 sentiments:
∙ Positive Sentiments: love, happiness, fun, enthusiasm and relief,
∙ Negative Sentiments: hate, anger, sadness, boredom and worry,
∙ Neutral Sentiments: surprise and neutral.
Each tweet can be attributed more than one sentiment, depend- ing on their own judgement. The 12 sentiments chosen are the most common ones present in the dataset. Tweets that are attributed both a positive and a negative sentiment are removed from the data set. In total, we obtained 18 153 positive tweets, 11 999 negative ones, and 5 626 neutral ones.
The tweets are divided into 3 sets as follows:
∙ Set 1: this dataset is used during our experiment to extract patterns related to each sentiment (This will be discussed in the next subsection). Therefore, in the rest of this work, we will refer to this set as the “Pattern Extraction set”.
∙ Set 2: this set is used for training. Therefore, in the rest of this work, it will be referred to as the “Training set”.
∙ Set 3: this set will serve as a test set. Therefore, in the rest of this work, it will be referred to as the “Test set”.
In TABLE I, we describe how the tweets are distributed among the 3 sets defined above.
C. Features Extraction
In [2], we used 4 sets of features to classify tweets into 7 different sentiment classes. In our current work, we opt for the same sets of features to perform the ternary classification. We also, used the “unigram-based features”, and the “pattern- based features”, to attribute sentiment scores to the different sentiment sub-classes defined above. Therefore, in this sub- section, we briefly explain how the different sets of features are extracted.
1http://help.sentiment140.com/for-students
Authorized licensed use limited to: University of the Cumberlands. Downloaded on July 24,2021 at 03:23:21 UTC from IEEE Xplore. Restrictions apply.
1) Sentiment-based features: Sentiment-based features are ones based on the sentiment polarity (i.e., positive/negative) of the different components of tweets. These features are extracted using SentiStrength 2. The features extracted are the following:
∙ Total score of positive words denoted as 𝑃𝑊 ∙ Total Score of negative words denoted as 𝑁𝑊 (this score
is positive) ∙ Number of highly emotional positive words (i.e., having
score equal to or more than 3) denoted as 𝑁𝑝𝑤 ∙ Number of highly emotional negative words (i.e., having
score equal to or less than -3) denoted as 𝑁𝑛𝑤 ∙ Ratio of emotional words 𝜌(𝑡) defined as
𝜌(𝑡) = 𝑃𝑊(𝑡) − 𝑁𝑊(𝑡) 𝑃𝑊(𝑡) + 𝑁𝑊(𝑡)
(1)
where 𝑡 is the tweet. In case the tweet does not contain any emotional word, 𝜌 is set to 0.
Six more features are extracted by counting the number of positive, negative, joking (or ironic), and neutral emoticons, as well as the number of positive and negative hashtags; and four features are extracted that represent whether there is a sentiment contrast between words, between hashtags, between words and hashtags, and between words and emoticons.
2) punctuation and syntax-based features: In addition to sentiment-based features, a second set of features qualified of punctuation-based features is extracted. This set includes:
∙ Number of exclamation marks ∙ Number of question marks ∙ Number of dots ∙ Number of all-capital words ∙ Number of quotes
A sixth feature is added by checking whether any of the words contains a vowel that is repeated more than twice (e.g., “looooove”). If such a word exists, the feature is set to “true”, otherwise, it is set to “false”.
3) Unigram-based features: These features are extracted using WordNet [11]. WordNet is used to collect unigrams related to each sentiment class. We start with a set of seed words few in number for each class (except the classes “neutral” and “surprise”), and used WordNet to collect their synonyms and hyponyms down to a certain depth.
The choice of synonyms and hyponyms is explained in de- tails in [2]. However, it is important to mention that hyponyms might lose the original meaning of the word, and collide with some of other classes. Therefore, the depth down to which we collect the hyponyms is set to a certain value 𝐷ℎ𝑦𝑝𝑜, which is a parameter to optimize. To decide the value of 𝐷ℎ𝑦𝑝𝑜, we rely on the manual check of the returned list of words at each depth, to decide on whether or not the words are deviating much from the original seed words meaning. In the rest of this work, 𝐷ℎ𝑦𝑝𝑜 is set to 2.
We use the resulted sets to extract 10 features, by counting the occurrences of the words in the tweet to classify.
2http://sentistrength.wlv.ac.uk/
TABLE II EXPRESSIONS USED TO REPLACE THE WORDS OF EW AND GFI
PoS Tag Expression “CC” COORDCONJUNCTION “CD” CARDINAL “DT” DETERMINER “EX” EXISTTHERE “FW” FOREIGNWORD “IN” PREPOSITION “LS” LISTMARKER “MD” MODAL “PDT” PREDETERMINER “POS” POSSESSIVEEND “PRP”, “PRP$” PRONOUN “RP” PARTICLE “SYM” SYMBOL “TO” TO “UH” INTERJECTION “WDT”, “WP”, “WP$”, “WRB” WHDETERMINER “NN”, “NNS”, “NNP”, “NNPS” NOUN
TABLE III PART-OF-SPEECH TAG CLASSES
Class PoS Tags
EW “CC, “DT”, “IN”, “MD”, “RP”, “TO”, “PRP”, “PRP$”, “NNP”, “NNPS”, “SYM”, “UH”, “EX”, “PDT”, “POS”, “WDT”, “WP”, “WP$”,“WRB”, “CD”, “FW”, “LS”
CI “VB, “VBD”, “VBG”, “VBN”, “VBP”, “VBZ”, “JJ”, “JJR”, “JJS”
GFI “NN, “NNS”, “RB”, “RBR”, “RBS”
4) Pattern-based features: In [2], the proposed approach relies on Part of Speech (PoS) tags to extract patterns for each sentiment class from the training set, and proposed a scoring function to measure the resemblance of tweets to patterns of different sentiments classes. In our work, we opt for a similar approach: we divide words into three classes: a first one referred to as EW containing words which might have emotional content, a second class referred to as “CI” containing words of which the content is important, and a third one referred to as “GFI” containing the words whose grammatical function is important. If a word belongs to the first category, it is replaced by the corresponding expression shown in TABLE II. If the word belongs to the second, it is lemmatized and replaced by its lemma. Otherwise, if it belongs to the third category, it is replaced it by the corresponding expression shown in TABLE II along with its sentiment polarity (i.e., positive, negative or neutral). The classification is done based on the PoS tag of the word in the tweet. The list of PoS tags and their classes is given in TABLE III.
We generate the vector of words for each tweet as described. We then define a pattern as an ordered sequence of words. The patterns are extracted from the pattern extraction set and are taken such as their length satisfies
𝐿𝑚𝑖𝑛 ≤ 𝐿𝑒𝑛𝑔𝑡ℎ(𝑝𝑎𝑡𝑡𝑒𝑟𝑛) ≤ 𝐿𝑚𝑎𝑥 (2) where 𝐿𝑚𝑖𝑛 and 𝐿𝑚𝑎𝑥 represent respectively the minimal and maximal allowed length of patterns in 𝑤𝑜𝑟𝑑𝑠. The number of pattern lengths is 𝑁𝐿 = (𝐿𝑚𝑎𝑥 − 𝐿𝑚𝑖𝑛 + 1).
Authorized licensed use limited to: University of the Cumberlands. Downloaded on July 24,2021 at 03:23:21 UTC from IEEE Xplore. Restrictions apply.
TABLE IV PATTERN FEATURES
Pattern length 𝐿1 𝐿2 ⋅ ⋅ ⋅ 𝐿𝑁𝐿
1 𝐹1,1 𝐹1,2 ⋅ ⋅ ⋅ 𝐹1,𝑁𝐿 Sentiment 2 𝐹2,1 𝐹2,2 ⋅ ⋅ ⋅ 𝐹2,𝑁𝐿
Class ...
... ...
. . . ...
12 𝐹12,1 𝐹12,2 ⋅ ⋅ ⋅ 𝐹12,𝑁𝐿
Only patterns that appear at least 𝑁𝑜𝑐𝑐 times in our training set are kept; the others are discarded. After the selection, we divide the resulted patterns into 𝑁𝐹 sets as follows:
𝑁𝐹 = 𝑁𝐿 × 𝑁𝐶 (3)
where 𝑁𝐿 the number of pattern lengths and 𝑁𝐶 is the number of classes (12 in our case). We create 𝑁𝐹 features, as shown in TABLE IV. Each feature 𝐹𝑖,𝑗 of the table represents the degree of resemblance of the tweet to the patterns of sentiment class 𝑖 and length 𝑗. Therefore, given a tweet 𝑡, we calculate the resemblance degree 𝑟𝑒𝑠(𝑝, 𝑡) of each pattern 𝑝 to the tweet 𝑡.
𝑟𝑒𝑠(𝑝, 𝑡) =
⎧ ⎨ ⎩
1, if the tweet vector contains the pattern as it is, in the same order,
𝛼 ⋅ 𝑛/𝑁, if 𝑛 words out of the 𝑁 words of the pattern appear in the tweet in the correct order,
0, if no word of the pattern appears in the tweet.
Given the 𝑘𝑛𝑛 patterns that have the highest resemblance to the pattern 𝑝 among the patterns extracted from the class 𝑖 which have a length 𝑗, the value of the feature 𝐹𝑖,𝑗 is
𝐹𝑖,𝑗 = 𝛽𝑗 ∗ 𝑘𝑛𝑛∑ 𝑘=1
𝑟𝑒𝑠(𝑝𝑘, 𝑡) (4)
where 𝛽𝑗 is a weight given to patterns of length 𝐿𝑗 (regardless of their class). We give different weights for each length of pattern since longer patterns are more likely to have higher impact.
In previous works [12][2], we demonstrated that the optimal values for the different parameters are as follows: ⎧ ⎨ ⎩
𝑁𝑜𝑐𝑐 = 2,
𝑘𝑛𝑛 = 5,
𝐿𝑚𝑖𝑛 = 3,
𝐿𝑚𝑎𝑥 = 10,
𝛼 = 0.03,
𝛽𝑛 = (𝑛 − 1)/(𝑛 + 1), ∀𝑛 ∈ {3, . . . , 10}. The sets of features as described are used to perform
the ternary classification and to attribute sentiment scores to tweets.
Fig. 1. Classification accuracy per sentiment and total classification for the different sets of features
IV. EXPERIMENTAL RESULTS
A. Ternary Classification
The classification is performed using Random Forest. In a first step, we examine the performance of classification per feature set. The results of classification are given in Fig. 1.
The ternary classification results show that all of the sets of features performed well. However, the sentiment-based features remarkably outperform the other sets when detect- ing negative tweets. This is because sentiment features, as proposed, were specifically designed to detect the polarity of the tweet (i.e., whether it is positive or negative), unlike the other sets of features which were proposed to measure the resemblance to the training data. This leads to the next observation that the classification of positive sentiment is quite high for all the sets of features. This classification accuracy is obtained thanks to the big enough training data for this class.
More interestingly, we notice that, individually, the sets of features cannot detect neutral statements well. However, if combined, the accuracy of classification of neutral tweets is better. Neutral tweets are hard to detect and identify. This is because the two other classes can be considered as two opposite poles while the neutral class is in between with a small coverage. Tweets usually tend to be ”polarized” (i.e., classified as sentimental). This can be explained by the following reasons:
∙ Unigram-based features do not contain a feature that is specific to non sentimental tweets.
∙ Pattern-based features return the degree of resemblance of a tweet to patterns extracted from each sentiment class. The neutral tweets represent a “minority” in the three sets used. Therefore, the amount of patterns extracted for this class, as well as the training data are small.
∙ The initial set of tweets was extracted from a big file containing only “positive” and “negative” tweets. The manual annotation that returned neutral class was non- expected.
The overall classification results are given in TABLE V. The obtained accuracy is close to the one in [2] which is 83.0%.
Authorized licensed use limited to: University of the Cumberlands. Downloaded on July 24,2021 at 03:23:21 UTC from IEEE Xplore. Restrictions apply.
TABLE V TERNARY CLASSIFICATION PERFORAMANCE
Class TP Rate FP Rate Precision Recall F-Measure Positive 92.1% 31.2% 81.3% 92.1% 86.4% Neutral 35.6% 2.1% 65.1% 35.6% 46.0% Negative 74.3% 6.4% 83.6% 74.3% 78.7% Total 81.0% 19.0% 81.0% 81.0% 81.0%
B. Quantification
To perform the quantification, we define the following scores for each tweet:
∙ Unigram-based score (𝑆𝑢): Given that 𝑁𝑖 unigrams of a sentiment class 𝑖 appear in the tweet 𝑡 as described in Section III, each unigram 𝑘 has its own score 𝑠𝑘 returned by SentiStrength. The unigram-based score of the tweet for the given class 𝑖 is defined as follows:
𝑆𝑢(𝑖) =
𝑁𝑖∑ 𝑘=1
𝑆𝑘. (5)
∙ Pattern-based score (𝑆𝑝): Given the 𝑘𝑛𝑛 patterns of length 𝑗 of a sentiment 𝑖, that resembles the most to the tweet’s patterns, and given the weights 𝛽𝑗 given to the patterns of length 𝑗. The pattern-based score of the tweet for the given sentiment class 𝑖 is defined as follows:
𝑆𝑝(𝑖) =
𝑁𝐿∑ 𝑗=1
𝛽𝑗 ⋅ 𝑘𝑛𝑛∑ 𝑘=1
𝑟𝑒𝑠(𝑝𝑘, 𝑡), (6)
where 𝑁𝐿 is the number of pattern lengths.
The total score of a sub-sentiment 𝑖 is:
𝑆(𝑖) = 𝜉 ⋅ 𝑆𝑢(𝑖) + (1 − 𝜉) ⋅ 𝑆𝑝(𝑖), (7) where 𝜉 is a weight such as
0 ≤ 𝜉 ≤ 1. (8) 1) Sentiment scores as validators of ternary classifica-
tion: In a first step, the two scores are used to validate the correctness of classification. Therefore, for the different values of 𝜉 = {0, 1/4, 1/2, 3/4, 1}, we sum both positive and negative sentiments scores (each polarity aside) and estimate the precision and coverage of classification.
Therefore, for each tweet judged as sentimental, we se- lect the positive/negative score returned. For each threshold {0, 1, ⋅ ⋅ ⋅ , 20}, we measure precision of classification of the tweets that have a score higher than the threshold, and the number of positive/negative tweets having such score over the total number of positive/negative tweets (i.e., coverage). The accuracy and coverage returned are given in Fig. 2.
The obtained results show that for a given precision re- quired, it is possible to set a threshold that satisfies the required precision, and returns the corresponding coverage. This result might be used to extract the most trustworthy data.
The results show that unigram-based scores are efficient to detect positive tweets with high precision: bigger values
TABLE VI DISTRIBUTION OF THE NUMBER OF SENTIMENTS PER TWEET IN EACH SET
Set 1 sent. 2 sent. 3 sent. 4 sent. Pattern Extraction set 1388 9309 5339 980 Training set 2199 6197 686 918 Test set 1575 5115 618 692
of 𝜉 present a faster convergence towards a precision equal to 100%. Pattern-based scores, on the other hand, are more efficient in detecting negative tweets with higher precision.
2) Sentiment scores as detectors of sentiment sub-classes: As mentioned above, the tweets are manually annotated. They are given one or more sentiments. TABLE VI shows the number of sentiments per tweet in the different sets:
Given that a tweet 𝑡 contains 𝑁 sentiments (according to the manual annotation), we collect the 𝑁 sentiments that have the highest scores and compare them to the ones given by the manual annotation. To measure the performances of our proposed scoring function as sentiment sub-class detectors, we define the two following metrics:
- Metric 1:
𝑀1(𝑡) =
⎧ ⎨ ⎩
0, if the ternary classification is wrong, or if it is correct but no sentiment sub-class is detected,
1, if the ternary classification is correct, and at least 1 out of the 𝑁 sentiments defined by the annotators is correctly detected.
This metric serves to verify the performance of detection of at least one sentiment present in the tweet.
- Metric 2:
𝑀2(𝑡) =
⎧ ⎨ ⎩
0, if the ternary classification is wrong, or if it is correct but no sentiment sub-class is detected,
𝑛/𝑁, if the tweet is sentimental, the ternary classification is correct, and 𝑛 out of the 𝑁 sentiments defined by the annotators are detected,
1, if the tweet is neutral, and is correctly classified.
This metric is stricter than 𝑀1. It serves to verify the performance of detection of all the sentiments present in the tweet.
The two metrics give the performances presented in TABLE VII. Obviously, the pattern-based scores serve as better senti- ment detectors, which goes along with the results of the work [2] that states that the pattern-based features have the highest contribution during the multi-class sentiment classification.
According to the results returned by the metric 𝑀1, it is always possible to detect a sentiment present in the tweet with a high accuracy. The sentiment detected that has the highest score, is very likely to be the one that is the most dominant.
Although the annotators have not given a specific order for the present sentiments, the quantification of sentiments,
Authorized licensed use limited to: University of the Cumberlands. Downloaded on July 24,2021 at 03:23:21 UTC from IEEE Xplore. Restrictions apply.
Fig. 2. Precision Vs Coverage for positive (Top) and negative (Bottom) tweets, for different values of 𝜉. (a) shows the results for 𝜉 = 1, (b) shows those for 𝜉 = 3/4, (c) shows those for 𝜉 = 1/2, (d) shows those for 𝜉 = 1/4, (e) shows those for 𝜉 = 0.
TABLE VII PERFORMANCES OF THE PROPOSED APPROACH ACCORDING TO THE DEFINED METRICS
Value of 𝜉 Metric 𝑀1 Metric 𝑀2 Positive Negative Neutral Overall Positive Negative Neutral Overall
1 52.5% 53.1% 65.2% 53.30% 29.0% 33.1% 65.2% 32.1% 3⁄4 62.1% 62.7% 65.2% 62.4% 41.3% 43.3% 65.2% 43.2% 1⁄2 64.8% 63.8% 65.2% 64.5% 44.3% 44.7% 65.2% 45.5% 1⁄4 65.6% 63.6% 65.2% 65.0% 45.6% 45.2% 65.2% 46.6% 0 65.9% 63.5% 65.2% 65.2% 46.1% 45.4% 65.2% 47.0%
implicitly allows giving an order of the existing sentiments. The metric 𝑀2 allows measuring the detection of the different existing sentiments in order. The value obtained (i.e. 47.0%) is relatively high for a very constrained metric. The result of this metric, compared to 𝑀1, reveals the importance of the other collected sentiments. As stated in [2], most of tweets that were misclassified, actually contained more than 1 sentiment.
V. CONCLUSION
In this paper, we proposed an approach for sentiment quantification in Twitter. This work highlighted the importance of the detection of all the sentiments in a tweet, and proposed a way to extract these different existing sentiments. For this sake, we proposed a set of pattern-based features and special unigram-based features along with the other basic features to perform the ternary classification, then the quantification of sentiments within tweets.In future work, we will work on the optimization of the parameters we defined during the extraction of pattern-based and unigram-based features/scores to obtain better performances.
ACKNOWLEDGMENT
The research results have been achieved by “Cognitive Security: A New Approach to Securing Future Large Scale and Distributed Mobile Applications,” the Commissioned Research of National Institute of Information and Communications Technology (NICT) , JAPAN.
REFERENCES
[1] K. Ghag and K. Shah, “Comparative analysis of the techniques for sentiment analysis,” in Proc. Int. Conf. Advances in Technology and Eng., pp. 1–7, Jan. 2013.
[2] M. Bouazizi and T. Ohtsuki, “Sentiment Analysis: from Binary to Multi-Class Classification - A Pattern-Based Approach for Multi-Class Sentiment Analysis in Twitter,” in Proc. IEEE ICC, May 2016.
[3] M. Boia, B. Faltings, C.-C. Musat and P. Pu, “A :) is worth a thousand words: How people attach sentiment to emoticons and words in tweets,” in Proc. Int. Conf. Social Computing, pp. 345–350, Sept. 2013.
[4] K. Manuel, K. V. Indukuri and P. R. Krishna, “Analyzing internet slang for sentiment mining,” in Proc. 2nd Vaagdevi Int. Conf. Inform. Technology for Real World Problems, pp. 9–11 Dec. 2010.
[5] M. Bouazizi and T. Ohtsuki, “Sentiment Analysis in Twitter for Multiple Topics - How to Detect the Polarity of Tweets Regardless of Their Topic,” in Proc. IEICE ASN, pp 91–96, Feb. 2015.
[6] J. M. Soler, F. Cuartero, and M. Roblizo, “Twitter as a tool for predicting elections results,” in Proc. IEEE/ACM ASONAM, pp. 1194–1200, Aug. 2012.
[7] H. A. Aldahawi and S. M. Allen, “Twitter mining in the oil business: A sentiment analysis approach,”in Proc. 3rd Int. Conf. Cloud and Green Computing, pp. 581–586, Sept.–Oct. 2013.
[8] K.H. Lin, C. Yang and H.Chen, “What emotions do news articles trigger in their readers?,” in Proc. ACM SIGIR ’07, pp. 733–734, July 2007.
[9] K.H. Lin, C. Yang and H Chen, “Emotion Classification of Online News Articles from the Reader’s Perspective,” in Proc. IEEE/WIC/ACM WI-IAT ’08, vol.1, pp.220–226, Dec. 2008.
[10] W. Liang, H. Wang, Y. Chu and C. Wu, “Emoticon recommendation in microblog using affective trajectory model,” in Proc. Annual Summit and Conf. Asia-Pacific Signal and Information Processing Association (APSIPA), pp.1–5, Dec. 2014.
[11] C. Fellbaun, WordNet: an Electronic Lexical Database, Cambridge, Massachusetts, 1998.
[12] M. Bouazizi and T. Ohtsuki, “Sarcasm Detection in Twitter,” in Proc. IEEE GLOBECOM,pp. 1–6 Dec. 2015.
Authorized licensed use limited to: University of the Cumberlands. Downloaded on July 24,2021 at 03:23:21 UTC from IEEE Xplore. Restrictions apply.
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (Gray Gamma 2.2) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.7 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 0 /ParseDSCComments false /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo false /PreserveFlatness true /PreserveHalftoneInfo true /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true /AbadiMT-CondensedLight /ACaslon-Italic /ACaslon-Regular /ACaslon-Semibold /ACaslon-SemiboldItalic /AdobeArabic-Bold /AdobeArabic-BoldItalic /AdobeArabic-Italic /AdobeArabic-Regular /AdobeHebrew-Bold /AdobeHebrew-BoldItalic /AdobeHebrew-Italic /AdobeHebrew-Regular /AdobeHeitiStd-Regular /AdobeMingStd-Light /AdobeMyungjoStd-Medium /AdobePiStd /AdobeSongStd-Light /AdobeThai-Bold /AdobeThai-BoldItalic /AdobeThai-Italic /AdobeThai-Regular /AGaramond-Bold /AGaramond-BoldItalic /AGaramond-Italic /AGaramond-Regular /AGaramond-Semibold /AGaramond-SemiboldItalic /AgencyFB-Bold /AgencyFB-Reg /AGOldFace-Outline /AharoniBold /Algerian /Americana /Americana-ExtraBold /AndaleMono /AndaleMonoIPA /AngsanaNew /AngsanaNew-Bold /AngsanaNew-BoldItalic /AngsanaNew-Italic /AngsanaUPC /AngsanaUPC-Bold /AngsanaUPC-BoldItalic /AngsanaUPC-Italic /Anna /ArialAlternative /ArialAlternativeSymbol /Arial-Black /Arial-BlackItalic /Arial-BoldItalicMT /Arial-BoldMT /Arial-ItalicMT /ArialMT /ArialMT-Black /ArialNarrow /ArialNarrow-Bold /ArialNarrow-BoldItalic /ArialNarrow-Italic /ArialRoundedMTBold /ArialUnicodeMS /ArrusBT-Bold /ArrusBT-BoldItalic /ArrusBT-Italic /ArrusBT-Roman /AvantGarde-Book /AvantGarde-BookOblique /AvantGarde-Demi /AvantGarde-DemiOblique /AvantGardeITCbyBT-Book /AvantGardeITCbyBT-BookOblique /BakerSignet /BankGothicBT-Medium /Barmeno-Bold /Barmeno-ExtraBold /Barmeno-Medium /Barmeno-Regular /Baskerville /BaskervilleBE-Italic /BaskervilleBE-Medium /BaskervilleBE-MediumItalic /BaskervilleBE-Regular /Baskerville-Bold /Baskerville-BoldItalic /Baskerville-Italic /BaskOldFace /Batang /BatangChe /Bauhaus93 /Bellevue /BellMT /BellMTBold /BellMTItalic /BerlingAntiqua-Bold /BerlingAntiqua-BoldItalic /BerlingAntiqua-Italic /BerlingAntiqua-Roman /BerlinSansFB-Bold /BerlinSansFBDemi-Bold /BerlinSansFB-Reg /BernardMT-Condensed /BernhardModernBT-Bold /BernhardModernBT-BoldItalic /BernhardModernBT-Italic /BernhardModernBT-Roman /BiffoMT /BinnerD /BinnerGothic /BlackadderITC-Regular /Blackoak /blex /blsy /Bodoni /Bodoni-Bold /Bodoni-BoldItalic /Bodoni-Italic /BodoniMT /BodoniMTBlack /BodoniMTBlack-Italic /BodoniMT-Bold /BodoniMT-BoldItalic /BodoniMTCondensed /BodoniMTCondensed-Bold /BodoniMTCondensed-BoldItalic /BodoniMTCondensed-Italic /BodoniMT-Italic /BodoniMTPosterCompressed /Bodoni-Poster /Bodoni-PosterCompressed /BookAntiqua /BookAntiqua-Bold /BookAntiqua-BoldItalic /BookAntiqua-Italic /Bookman-Demi /Bookman-DemiItalic /Bookman-Light /Bookman-LightItalic /BookmanOldStyle /BookmanOldStyle-Bold /BookmanOldStyle-BoldItalic /BookmanOldStyle-Italic /BookshelfSymbolOne-Regular /BookshelfSymbolSeven /BookshelfSymbolThree-Regular /BookshelfSymbolTwo-Regular /Botanical /Boton-Italic /Boton-Medium /Boton-MediumItalic /Boton-Regular /Boulevard /BradleyHandITC /Braggadocio /BritannicBold /Broadway /BrowalliaNew /BrowalliaNew-Bold /BrowalliaNew-BoldItalic /BrowalliaNew-Italic /BrowalliaUPC /BrowalliaUPC-Bold /BrowalliaUPC-BoldItalic /BrowalliaUPC-Italic /BrushScript /BrushScriptMT /CaflischScript-Bold /CaflischScript-Regular /Calibri /Calibri-Bold /Calibri-BoldItalic /Calibri-Italic /CalifornianFB-Bold /CalifornianFB-Italic /CalifornianFB-Reg /CalisMTBol /CalistoMT /CalistoMT-BoldItalic /CalistoMT-Italic /Cambria /Cambria-Bold /Cambria-BoldItalic /Cambria-Italic /CambriaMath /Candara /Candara-Bold /Candara-BoldItalic /Candara-Italic /Carta /CaslonOpenfaceBT-Regular /Castellar /CastellarMT /Centaur /Centaur-Italic /Century /CenturyGothic /CenturyGothic-Bold /CenturyGothic-BoldItalic /CenturyGothic-Italic /CenturySchL-Bold /CenturySchL-BoldItal /CenturySchL-Ital /CenturySchL-Roma /CenturySchoolbook /CenturySchoolbook-Bold /CenturySchoolbook-BoldItalic /CenturySchoolbook-Italic /CGTimes-Bold /CGTimes-BoldItalic /CGTimes-Italic /CGTimes-Regular /CharterBT-Bold /CharterBT-BoldItalic /CharterBT-Italic /CharterBT-Roman /CheltenhamITCbyBT-Bold /CheltenhamITCbyBT-BoldItalic /CheltenhamITCbyBT-Book /CheltenhamITCbyBT-BookItalic /Chiller-Regular /Cmb10 /CMB10 /Cmbsy10 /CMBSY10 /CMBSY5 /CMBSY6 /CMBSY7 /CMBSY8 /CMBSY9 /Cmbx10 /CMBX10 /Cmbx12 /CMBX12 /Cmbx5 /CMBX5 /Cmbx6 /CMBX6 /Cmbx7 /CMBX7 /Cmbx8 /CMBX8 /Cmbx9 /CMBX9 /Cmbxsl10 /CMBXSL10 /Cmbxti10 /CMBXTI10 /Cmcsc10 /CMCSC10 /Cmcsc8 /CMCSC8 /Cmcsc9 /CMCSC9 /Cmdunh10 /CMDUNH10 /Cmex10 /CMEX10 /CMEX7 /CMEX8 /CMEX9 /Cmff10 /CMFF10 /Cmfi10 /CMFI10 /Cmfib8 /CMFIB8 /Cminch /CMINCH /Cmitt10 /CMITT10 /Cmmi10 /CMMI10 /Cmmi12 /CMMI12 /Cmmi5 /CMMI5 /Cmmi6 /CMMI6 /Cmmi7 /CMMI7 /Cmmi8 /CMMI8 /Cmmi9 /CMMI9 /Cmmib10 /CMMIB10 /CMMIB5 /CMMIB6 /CMMIB7 /CMMIB8 /CMMIB9 /Cmr10 /CMR10 /Cmr12 /CMR12 /Cmr17 /CMR17 /Cmr5 /CMR5 /Cmr6 /CMR6 /Cmr7 /CMR7 /Cmr8 /CMR8 /Cmr9 /CMR9 /Cmsl10 /CMSL10 /Cmsl12 /CMSL12 /Cmsl8 /CMSL8 /Cmsl9 /CMSL9 /Cmsltt10 /CMSLTT10 /Cmss10 /CMSS10 /Cmss12 /CMSS12 /Cmss17 /CMSS17 /Cmss8 /CMSS8 /Cmss9 /CMSS9 /Cmssbx10 /CMSSBX10 /Cmssdc10 /CMSSDC10 /Cmssi10 /CMSSI10 /Cmssi12 /CMSSI12 /Cmssi17 /CMSSI17 /Cmssi8 /CMSSI8 /Cmssi9 /CMSSI9 /Cmssq8 /CMSSQ8 /Cmssqi8 /CMSSQI8 /Cmsy10 /CMSY10 /Cmsy5 /CMSY5 /Cmsy6 /CMSY6 /Cmsy7 /CMSY7 /Cmsy8 /CMSY8 /Cmsy9 /CMSY9 /Cmtcsc10 /CMTCSC10 /Cmtex10 /CMTEX10 /Cmtex8 /CMTEX8 /Cmtex9 /CMTEX9 /Cmti10 /CMTI10 /Cmti12 /CMTI12 /Cmti7 /CMTI7 /Cmti8 /CMTI8 /Cmti9 /CMTI9 /Cmtt10 /CMTT10 /Cmtt12 /CMTT12 /Cmtt8 /CMTT8 /Cmtt9 /CMTT9 /Cmu10 /CMU10 /Cmvtt10 /CMVTT10 /ColonnaMT /Colossalis-Bold /ComicSansMS /ComicSansMS-Bold /Consolas /Consolas-Bold /Consolas-BoldItalic /Consolas-Italic /Constantia /Constantia-Bold /Constantia-BoldItalic /Constantia-Italic /CooperBlack /CopperplateGothic-Bold /CopperplateGothic-Light /Copperplate-ThirtyThreeBC /Corbel /Corbel-Bold /Corbel-BoldItalic /Corbel-Italic /CordiaNew /CordiaNew-Bold /CordiaNew-BoldItalic /CordiaNew-Italic /CordiaUPC /CordiaUPC-Bold /CordiaUPC-BoldItalic /CordiaUPC-Italic /Courier /Courier-Bold /Courier-BoldOblique /CourierNewPS-BoldItalicMT /CourierNewPS-BoldMT /CourierNewPS-ItalicMT /CourierNewPSMT /Courier-Oblique /CourierStd /CourierStd-Bold /CourierStd-BoldOblique /CourierStd-Oblique /CourierX-Bold /CourierX-BoldOblique /CourierX-Oblique /CourierX-Regular /CreepyRegular /CurlzMT /David-Bold /David-Reg /DavidTransparent /Dcb10 /Dcbx10 /Dcbxsl10 /Dcbxti10 /Dccsc10 /Dcitt10 /Dcr10 /Desdemona /DilleniaUPC /DilleniaUPCBold /DilleniaUPCBoldItalic /DilleniaUPCItalic /Dingbats /DomCasual /Dotum /DotumChe /DoulosSIL /EdwardianScriptITC /Elephant-Italic /Elephant-Regular /EngraversGothicBT-Regular /EngraversMT /EraserDust /ErasITC-Bold /ErasITC-Demi /ErasITC-Light /ErasITC-Medium /ErieBlackPSMT /ErieLightPSMT /EriePSMT /EstrangeloEdessa /Euclid /Euclid-Bold /Euclid-BoldItalic /EuclidExtra /EuclidExtra-Bold /EuclidFraktur /EuclidFraktur-Bold /Euclid-Italic /EuclidMathOne /EuclidMathOne-Bold /EuclidMathTwo /EuclidMathTwo-Bold /EuclidSymbol /EuclidSymbol-Bold /EuclidSymbol-BoldItalic /EuclidSymbol-Italic /EucrosiaUPC /EucrosiaUPCBold /EucrosiaUPCBoldItalic /EucrosiaUPCItalic /EUEX10 /EUEX7 /EUEX8 /EUEX9 /EUFB10 /EUFB5 /EUFB7 /EUFM10 /EUFM5 /EUFM7 /EURB10 /EURB5 /EURB7 /EURM10 /EURM5 /EURM7 /EuroMono-Bold /EuroMono-BoldItalic /EuroMono-Italic /EuroMono-Regular /EuroSans-Bold /EuroSans-BoldItalic /EuroSans-Italic /EuroSans-Regular /EuroSerif-Bold /EuroSerif-BoldItalic /EuroSerif-Italic /EuroSerif-Regular /EUSB10 /EUSB5 /EUSB7 /EUSM10 /EUSM5 /EUSM7 /FelixTitlingMT /Fences /FencesPlain /FigaroMT /FixedMiriamTransparent /FootlightMTLight /Formata-Italic /Formata-Medium /Formata-MediumItalic /Formata-Regular /ForteMT /FranklinGothic-Book /FranklinGothic-BookItalic /FranklinGothic-Demi /FranklinGothic-DemiCond /FranklinGothic-DemiItalic /FranklinGothic-Heavy /FranklinGothic-HeavyItalic /FranklinGothicITCbyBT-Book /FranklinGothicITCbyBT-BookItal /FranklinGothicITCbyBT-Demi /FranklinGothicITCbyBT-DemiItal /FranklinGothic-Medium /FranklinGothic-MediumCond /FranklinGothic-MediumItalic /FrankRuehl /FreesiaUPC /FreesiaUPCBold /FreesiaUPCBoldItalic /FreesiaUPCItalic /FreestyleScript-Regular /FrenchScriptMT /Frutiger-Black /Frutiger-BlackCn /Frutiger-BlackItalic /Frutiger-Bold /Frutiger-BoldCn /Frutiger-BoldItalic /Frutiger-Cn /Frutiger-ExtraBlackCn /Frutiger-Italic /Frutiger-Light /Frutiger-LightCn /Frutiger-LightItalic /Frutiger-Roman /Frutiger-UltraBlack /Futura-Bold /Futura-BoldOblique /Futura-Book /Futura-BookOblique /FuturaBT-Bold /FuturaBT-BoldItalic /FuturaBT-Book /FuturaBT-BookItalic /FuturaBT-Medium /FuturaBT-MediumItalic /Futura-Light /Futura-LightOblique /GalliardITCbyBT-Bold /GalliardITCbyBT-BoldItalic /GalliardITCbyBT-Italic /GalliardITCbyBT-Roman /Garamond /Garamond-Bold /Garamond-BoldCondensed /Garamond-BoldCondensedItalic /Garamond-BoldItalic /Garamond-BookCondensed /Garamond-BookCondensedItalic /Garamond-Italic /Garamond-LightCondensed /Garamond-LightCondensedItalic /Gautami /GeometricSlab703BT-Light /GeometricSlab703BT-LightItalic /Georgia /Georgia-Bold /Georgia-BoldItalic /Georgia-Italic /GeorgiaRef /Giddyup /Giddyup-Thangs /Gigi-Regular /GillSans /GillSans-Bold /GillSans-BoldItalic /GillSans-Condensed /GillSans-CondensedBold /GillSans-Italic /GillSans-Light /GillSans-LightItalic /GillSansMT /GillSansMT-Bold /GillSansMT-BoldItalic /GillSansMT-Condensed /GillSansMT-ExtraCondensedBold /GillSansMT-Italic /GillSans-UltraBold /GillSans-UltraBoldCondensed /GloucesterMT-ExtraCondensed /Gothic-Thirteen /GoudyOldStyleBT-Bold /GoudyOldStyleBT-BoldItalic /GoudyOldStyleBT-Italic /GoudyOldStyleBT-Roman /GoudyOldStyleT-Bold /GoudyOldStyleT-Italic /GoudyOldStyleT-Regular /GoudyStout /GoudyTextMT-LombardicCapitals /GSIDefaultSymbols /Gulim /GulimChe /Gungsuh /GungsuhChe /Haettenschweiler /HarlowSolid /Harrington /Helvetica /Helvetica-Black /Helvetica-BlackOblique /Helvetica-Bold /Helvetica-BoldOblique /Helvetica-Condensed /Helvetica-Condensed-Black /Helvetica-Condensed-BlackObl /Helvetica-Condensed-Bold /Helvetica-Condensed-BoldObl /Helvetica-Condensed-Light /Helvetica-Condensed-LightObl /Helvetica-Condensed-Oblique /Helvetica-Fraction /Helvetica-Narrow /Helvetica-Narrow-Bold /Helvetica-Narrow-BoldOblique /Helvetica-Narrow-Oblique /Helvetica-Oblique /HighTowerText-Italic /HighTowerText-Reg /Humanist521BT-BoldCondensed /Humanist521BT-Light /Humanist521BT-LightItalic /Humanist521BT-RomanCondensed /Imago-ExtraBold /Impact /ImprintMT-Shadow /InformalRoman-Regular /IrisUPC /IrisUPCBold /IrisUPCBoldItalic /IrisUPCItalic /Ironwood /ItcEras-Medium /ItcKabel-Bold /ItcKabel-Book /ItcKabel-Demi /ItcKabel-Medium /ItcKabel-Ultra /JasmineUPC /JasmineUPC-Bold /JasmineUPC-BoldItalic /JasmineUPC-Italic /JoannaMT /JoannaMT-Italic /Jokerman-Regular /JuiceITC-Regular /Kartika /Kaufmann /KaufmannBT-Bold /KaufmannBT-Regular /KidTYPEPaint /KinoMT /KodchiangUPC /KodchiangUPC-Bold /KodchiangUPC-BoldItalic /KodchiangUPC-Italic /KorinnaITCbyBT-Regular /KristenITC-Regular /KrutiDev040Bold /KrutiDev040BoldItalic /KrutiDev040Condensed /KrutiDev040Italic /KrutiDev040Thin /KrutiDev040Wide /KrutiDev060 /KrutiDev060Bold /KrutiDev060BoldItalic /KrutiDev060Condensed /KrutiDev060Italic /KrutiDev060Thin /KrutiDev060Wide /KrutiDev070 /KrutiDev070Condensed /KrutiDev070Italic /KrutiDev070Thin /KrutiDev070Wide /KrutiDev080 /KrutiDev080Condensed /KrutiDev080Italic /KrutiDev080Wide /KrutiDev090 /KrutiDev090Bold /KrutiDev090BoldItalic /KrutiDev090Condensed /KrutiDev090Italic /KrutiDev090Thin /KrutiDev090Wide /KrutiDev100 /KrutiDev100Bold /KrutiDev100BoldItalic /KrutiDev100Condensed /KrutiDev100Italic /KrutiDev100Thin /KrutiDev100Wide /KrutiDev120 /KrutiDev120Condensed /KrutiDev120Thin /KrutiDev120Wide /KrutiDev130 /KrutiDev130Condensed /KrutiDev130Thin /KrutiDev130Wide /KunstlerScript /Latha /LatinWide /LetterGothic /LetterGothic-Bold /LetterGothic-BoldOblique /LetterGothic-BoldSlanted /LetterGothicMT /LetterGothicMT-Bold /LetterGothicMT-BoldOblique /LetterGothicMT-Oblique /LetterGothic-Slanted /LevenimMT /LevenimMTBold /LilyUPC /LilyUPCBold /LilyUPCBoldItalic /LilyUPCItalic /Lithos-Black /Lithos-Regular /LotusWPBox-Roman /LotusWPIcon-Roman /LotusWPIntA-Roman /LotusWPIntB-Roman /LotusWPType-Roman /LucidaBright /LucidaBright-Demi /LucidaBright-DemiItalic /LucidaBright-Italic /LucidaCalligraphy-Italic /LucidaConsole /LucidaFax /LucidaFax-Demi /LucidaFax-DemiItalic /LucidaFax-Italic /LucidaHandwriting-Italic /LucidaSans /LucidaSans-Demi /LucidaSans-DemiItalic /LucidaSans-Italic /LucidaSans-Typewriter /LucidaSans-TypewriterBold /LucidaSans-TypewriterBoldOblique /LucidaSans-TypewriterOblique /LucidaSansUnicode /Lydian /Magneto-Bold /MaiandraGD-Regular /Mangal-Regular /Map-Symbols /MathA /MathB /MathC /Mathematica1 /Mathematica1-Bold /Mathematica1Mono /Mathematica1Mono-Bold /Mathematica2 /Mathematica2-Bold /Mathematica2Mono /Mathematica2Mono-Bold /Mathematica3 /Mathematica3-Bold /Mathematica3Mono /Mathematica3Mono-Bold /Mathematica4 /Mathematica4-Bold /Mathematica4Mono /Mathematica4Mono-Bold /Mathematica5 /Mathematica5-Bold /Mathematica5Mono /Mathematica5Mono-Bold /Mathematica6 /Mathematica6Bold /Mathematica6Mono /Mathematica6MonoBold /Mathematica7 /Mathematica7Bold /Mathematica7Mono /Mathematica7MonoBold /MatisseITC-Regular /MaturaMTScriptCapitals /Mesquite /Mezz-Black /Mezz-Regular /MICR /MicrosoftSansSerif /MingLiU /Minion-BoldCondensed /Minion-BoldCondensedItalic /Minion-Condensed /Minion-CondensedItalic /Minion-Ornaments /MinionPro-Bold /MinionPro-BoldIt /MinionPro-It /MinionPro-Regular /Miriam /MiriamFixed /MiriamTransparent /Mistral /Modern-Regular /MonotypeCorsiva /MonotypeSorts /MSAM10 /MSAM5 /MSAM6 /MSAM7 /MSAM8 /MSAM9 /MSBM10 /MSBM5 /MSBM6 /MSBM7 /MSBM8 /MSBM9 /MS-Gothic /MSHei /MSLineDrawPSMT /MS-Mincho /MSOutlook /MS-PGothic /MS-PMincho /MSReference1 /MSReference2 /MSReferenceSansSerif /MSReferenceSansSerif-Bold /MSReferenceSansSerif-BoldItalic /MSReferenceSansSerif-Italic /MSReferenceSerif /MSReferenceSerif-Bold /MSReferenceSerif-BoldItalic /MSReferenceSerif-Italic /MSReferenceSpecialty /MSSong /MS-UIGothic /MT-Extra /MTExtraTiger /MT-Symbol /MT-Symbol-Italic /MVBoli /Myriad-Bold /Myriad-BoldItalic /Myriad-Italic /Myriad-Roman /Narkisim /NewCenturySchlbk-Bold /NewCenturySchlbk-BoldItalic /NewCenturySchlbk-Italic /NewCenturySchlbk-Roman /NewMilleniumSchlbk-BoldItalicSH /NewsGothic /NewsGothic-Bold /NewsGothicBT-Bold /NewsGothicBT-BoldItalic /NewsGothicBT-Italic /NewsGothicBT-Roman /NewsGothic-Condensed /NewsGothic-Italic /NewsGothicMT /NewsGothicMT-Bold /NewsGothicMT-Italic /NiagaraEngraved-Reg /NiagaraSolid-Reg /NimbusMonL-Bold /NimbusMonL-BoldObli /NimbusMonL-Regu /NimbusMonL-ReguObli /NimbusRomNo9L-Medi /NimbusRomNo9L-MediItal /NimbusRomNo9L-Regu /NimbusRomNo9L-ReguItal /NimbusSanL-Bold /NimbusSanL-BoldCond /NimbusSanL-BoldCondItal /NimbusSanL-BoldItal /NimbusSanL-Regu /NimbusSanL-ReguCond /NimbusSanL-ReguCondItal /NimbusSanL-ReguItal /Nimrod /Nimrod-Bold /Nimrod-BoldItalic /Nimrod-Italic /NSimSun /Nueva-BoldExtended /Nueva-BoldExtendedItalic /Nueva-Italic /Nueva-Roman /NuptialScript /OCRA /OCRA-Alternate /OCRAExtended /OCRB /OCRB-Alternate /OfficinaSans-Bold /OfficinaSans-BoldItalic /OfficinaSans-Book /OfficinaSans-BookItalic /OfficinaSerif-Bold /OfficinaSerif-BoldItalic /OfficinaSerif-Book /OfficinaSerif-BookItalic /OldEnglishTextMT /Onyx /OnyxBT-Regular /OzHandicraftBT-Roman /PalaceScriptMT /Palatino-Bold /Palatino-BoldItalic /Palatino-Italic /PalatinoLinotype-Bold /PalatinoLinotype-BoldItalic /PalatinoLinotype-Italic /PalatinoLinotype-Roman /Palatino-Roman /PapyrusPlain /Papyrus-Regular /Parchment-Regular /Parisian /ParkAvenue /Penumbra-SemiboldFlare /Penumbra-SemiboldSans /Penumbra-SemiboldSerif /PepitaMT /Perpetua /Perpetua-Bold /Perpetua-BoldItalic /Perpetua-Italic /PerpetuaTitlingMT-Bold /PerpetuaTitlingMT-Light /PhotinaCasualBlack /Playbill /PMingLiU /Poetica-SuppOrnaments /PoorRichard-Regular /PopplLaudatio-Italic /PopplLaudatio-Medium /PopplLaudatio-MediumItalic /PopplLaudatio-Regular /PrestigeElite /Pristina-Regular /PTBarnumBT-Regular /Raavi /RageItalic /Ravie /RefSpecialty /Ribbon131BT-Bold /Rockwell /Rockwell-Bold /Rockwell-BoldItalic /Rockwell-Condensed /Rockwell-CondensedBold /Rockwell-ExtraBold /Rockwell-Italic /Rockwell-Light /Rockwell-LightItalic /Rod /RodTransparent /RunicMT-Condensed /Sanvito-Light /Sanvito-Roman /ScriptC /ScriptMTBold /SegoeUI /SegoeUI-Bold /SegoeUI-BoldItalic /SegoeUI-Italic /Serpentine-BoldOblique /ShelleyVolanteBT-Regular /ShowcardGothic-Reg /Shruti /SILDoulosIPA /SimHei /SimSun /SimSun-PUA /SnapITC-Regular /StandardSymL /Stencil /StoneSans /StoneSans-Bold /StoneSans-BoldItalic /StoneSans-Italic /StoneSans-Semibold /StoneSans-SemiboldItalic /Stop /Swiss721BT-BlackExtended /Sylfaen /Symbol /SymbolMT /SymbolTiger /SymbolTigerExpert /Tahoma /Tahoma-Bold /Tci1 /Tci1Bold /Tci1BoldItalic /Tci1Italic /Tci2 /Tci2Bold /Tci2BoldItalic /Tci2Italic /Tci3 /Tci3Bold /Tci3BoldItalic /Tci3Italic /Tci4 /Tci4Bold /Tci4BoldItalic /Tci4Italic /TechnicalItalic /TechnicalPlain /Tekton /Tekton-Bold /TektonMM /Tempo-HeavyCondensed /Tempo-HeavyCondensedItalic /TempusSansITC /Tiger /TigerExpert /Times-Bold /Times-BoldItalic /Times-BoldItalicOsF /Times-BoldSC /Times-ExtraBold /Times-Italic /Times-ItalicOsF /TimesNewRomanMT-ExtraBold /TimesNewRomanPS-BoldItalicMT /TimesNewRomanPS-BoldMT /TimesNewRomanPS-ItalicMT /TimesNewRomanPSMT /Times-Roman /Times-RomanSC /Trajan-Bold /Trebuchet-BoldItalic /TrebuchetMS /TrebuchetMS-Bold /TrebuchetMS-Italic /Tunga-Regular /TwCenMT-Bold /TwCenMT-BoldItalic /TwCenMT-Condensed /TwCenMT-CondensedBold /TwCenMT-CondensedExtraBold /TwCenMT-CondensedMedium /TwCenMT-Italic /TwCenMT-Regular /Univers-Bold /Univers-BoldItalic /UniversCondensed-Bold /UniversCondensed-BoldItalic /UniversCondensed-Medium /UniversCondensed-MediumItalic /Univers-Medium /Univers-MediumItalic /URWBookmanL-DemiBold /URWBookmanL-DemiBoldItal /URWBookmanL-Ligh /URWBookmanL-LighItal /URWChanceryL-MediItal /URWGothicL-Book /URWGothicL-BookObli /URWGothicL-Demi /URWGothicL-DemiObli /URWPalladioL-Bold /URWPalladioL-BoldItal /URWPalladioL-Ital /URWPalladioL-Roma /USPSBarCode /VAGRounded-Black /VAGRounded-Bold /VAGRounded-Light /VAGRounded-Thin /Verdana /Verdana-Bold /Verdana-BoldItalic /Verdana-Italic /VerdanaRef /VinerHandITC /Viva-BoldExtraExtended /Vivaldii /Viva-LightCondensed /Viva-Regular /VladimirScript /Vrinda /Webdings /Westminster /Willow /Wingdings2 /Wingdings3 /Wingdings-Regular /WNCYB10 /WNCYI10 /WNCYR10 /WNCYSC10 /WNCYSS10 /WoodtypeOrnaments-One /WoodtypeOrnaments-Two /WP-ArabicScriptSihafa /WP-ArabicSihafa /WP-BoxDrawing /WP-CyrillicA /WP-CyrillicB /WP-GreekCentury /WP-GreekCourier /WP-GreekHelve /WP-HebrewDavid /WP-IconicSymbolsA /WP-IconicSymbolsB /WP-Japanese /WP-MathA /WP-MathB /WP-MathExtendedA /WP-MathExtendedB /WP-MultinationalAHelve /WP-MultinationalARoman /WP-MultinationalBCourier /WP-MultinationalBHelve /WP-MultinationalBRoman /WP-MultinationalCourier /WP-Phonetic /WPTypographicSymbols /XYATIP10 /XYBSQL10 /XYBTIP10 /XYCIRC10 /XYCMAT10 /XYCMBT10 /XYDASH10 /XYEUAT10 /XYEUBT10 /ZapfChancery-MediumItalic /ZapfDingbats /ZapfHumanist601BT-Bold /ZapfHumanist601BT-BoldItalic /ZapfHumanist601BT-Demi /ZapfHumanist601BT-DemiItalic /ZapfHumanist601BT-Italic /ZapfHumanist601BT-Roman /ZWAdobeF ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 2.00333 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /ColorImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 15 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 15 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 2.00333 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /GrayImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 15 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 15 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 600 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.00167 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false /Description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f7f75288fd94e9b8bbe5b9a521b5efa7684002000410064006f006200650020005000440046002065876863900275284e8e55464e1a65876863768467e5770b548c62535370300260a853ef4ee54f7f75280020004100630072006f0062006100740020548c002000410064006f00620065002000520065006100640065007200200035002e003000204ee553ca66f49ad87248672c676562535f00521b5efa768400200050004400460020658768633002> /CHT <FEFF4f7f752890194e9b8a2d7f6e5efa7acb7684002000410064006f006200650020005000440046002065874ef69069752865bc666e901a554652d965874ef6768467e5770b548c52175370300260a853ef4ee54f7f75280020004100630072006f0062006100740020548c002000410064006f00620065002000520065006100640065007200200035002e003000204ee553ca66f49ad87248672c4f86958b555f5df25efa7acb76840020005000440046002065874ef63002> /CZE <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> /DAN <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> /DEU <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> /ESP <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> /FRA <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> /GRE <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a stvaranje Adobe PDF dokumenata pogodnih za pouzdani prikaz i ispis poslovnih dokumenata koristite ove postavke. Stvoreni PDF dokumenti mogu se otvoriti Acrobat i Adobe Reader 5.0 i kasnijim verzijama.) /HUN <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> /ITA (Utilizzare queste impostazioni per creare documenti Adobe PDF adatti per visualizzare e stampare documenti aziendali in modo affidabile. I documenti PDF creati possono essere aperti con Acrobat e Adobe Reader 5.0 e versioni successive.) /JPN <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> /KOR <FEFFc7740020c124c815c7440020c0acc6a9d558c5ec0020be44c988b2c8c2a40020bb38c11cb97c0020c548c815c801c73cb85c0020bcf4ace00020c778c1c4d558b2940020b3700020ac00c7a50020c801d569d55c002000410064006f0062006500200050004400460020bb38c11cb97c0020c791c131d569b2c8b2e4002e0020c774b807ac8c0020c791c131b41c00200050004400460020bb38c11cb2940020004100630072006f0062006100740020bc0f002000410064006f00620065002000520065006100640065007200200035002e00300020c774c0c1c5d0c11c0020c5f40020c2180020c788c2b5b2c8b2e4002e> /NLD (Gebruik deze instellingen om Adobe PDF-documenten te maken waarmee zakelijke documenten betrouwbaar kunnen worden weergegeven en afgedrukt. De gemaakte PDF-documenten kunnen worden geopend met Acrobat en Adobe Reader 5.0 en hoger.) /NOR <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> /POL <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> /PTB <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> /RUM <FEFF005500740069006c0069007a00610163006900200061006300650073007400650020007300650074010300720069002000700065006e007400720075002000610020006300720065006100200064006f00630075006d0065006e00740065002000410064006f006200650020005000440046002000610064006500630076006100740065002000700065006e007400720075002000760069007a00750061006c0069007a00610072006500610020015f006900200074006900700103007200690072006500610020006c0061002000630061006c006900740061007400650020007300750070006500720069006f0061007201030020006100200064006f00630075006d0065006e00740065006c006f007200200064006500200061006600610063006500720069002e002000200044006f00630075006d0065006e00740065006c00650020005000440046002000630072006500610074006500200070006f00740020006600690020006400650073006300680069007300650020006300750020004100630072006f006200610074002c002000410064006f00620065002000520065006100640065007200200035002e00300020015f00690020007600650072007300690075006e0069006c006500200075006c0074006500720069006f006100720065002e> /RUS <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> /SLV <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> /SUO <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> /SVE <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> /TUR <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> /ENU (Use these settings to create Adobe PDF documents suitable for reliable viewing and printing of business documents. Created PDF documents can be opened with Acrobat and Adobe Reader 5.0 and later.) >> >> setdistillerparams << /HWResolution [600 600] /PageSize [612.000 792.000] >> setpagedevice
