Part 4 BEHS Presentation Project

Reference 1 – Cross-cultural reference

Citation in APA format (5 pts)

Wu, X., and Zhang, X., (2016) Automated Inference on Criminality using Face Images. arXiv:1611.04135v1 [cs.CV] 13 Nov 2016.

Key findings (10 pts)

In 2-3 paragraphs, summarize the main findings in your source.

Wu and Zhang studied the relation of facial appearance to criminology. Machine learning algorithms of Support Vector Machine, K-Nearest Neighbor, Convolutional Neural Network, and logistic regression were used to study 1856 facial images of real people. Half of the data used was obtained from convicted criminal records while the rest was of people who had no history of criminology. Additionally, the researchers balanced issues of age, gender, race, and different facial expressions.

The researchers used graphical methods to evaluate the accuracy of the four classifiers. CNN was found to have the highest accuracy of 89.51% on the validity of the automated face-induced inference on criminology. It was observed that supervised machine learning can identify criminal faces from the population using still images. However, this was the first research of its kind so more study is necessary to validate its usability.

How do you know that this is a credible/scholarly source? (5 pts)

Xiaolin Wu and Xi Zhang are both engineering researches at the Shanghai Jiao Tong University in China. Publication of their paper in Researchgate and arXvi repositories is proof of the credibility of the research.

Reference 2 – Policy reference

Citation in APA format (5 pts)

Internet Society, (2017, April 24). Artificial Intelligence and Machine Learning: Policy Paper

Key findings (10 pts)

In 2-3 paragraphs, summarize the main findings in your source.

The paper addresses key policies that should be considered when implementing Artificial Intelligence technologies. The policies provided are mainly that have a social impact on the use of the internet.

Internet Society requires researchers to maintain privacy for the data obtained during data mining. Researchers are further urged to showcase ethical practices in their innovations as some research may affect social relationships. Creation of artificial intelligence means the emergence of new ecosystems such as those of speech and smart agents. Users are required to keep these applications accessible and open to all.

Applications of artificial intelligence are also expected to have deep socio-economical impacts. The ability to learn, process natural language and plan may impact the social setup. Researchers should address all arising issues amicably.

How do you know that this is a credible/scholarly source? (5 pts)

Internet Society is a leader in setting Internet standards since 1992. The non-profit organization provides support for the standard-setting process with key participation in policies pertaining to cybersecurity.

Reference 3

Citation in APA format (5 pts)

Wang, T., Rudin, C., Wagner, D., & Sevieri, R. (2013, January). Detecting Patterns of Crime with Series Finder. In AAAI (Late-Breaking Developments).

Key findings (10 pts)

In 2-3 paragraphs, summarize the main findings in your source.

Crime analysts have to find patterns of ongoing crimes to determine where the next crime might occur. Series Finder is an algorithm that uses databases of past crimes to make this prediction. Series Finder identifies the modus operandi of the group or individual involved in crime and then narrows down to provide exact details of the offender.

Series Finder is based on previous crime analytic techniques such as association rule mining, classification, clustering, and pattern detection. The approach requires several pointers for it to predict correctly. High density cities and localized near repeats are examples of where this technology can be applied. Using simulated data, the researchers found Series Finder to predict possible crime occurrences with high precision.

How do you know that this is a credible/scholarly source? (5 pts)

Wang, Wagner, Sevieri, and Rudin were students at the Massachusetts Institute of Technology, Cambridge in the United States at the time of conducting this research. This paper was presented to the Twenty- Seventh AAAI Conference on Artificial Intelligence.

Reference 4

Citation in APA format (5 pts)

Saeed, U., Sarim, M., Usmani, A., Mukhtar, A., Shaikh, A., & Raffat, S. (2015). Application of machine learning algorithms in crime classification and classification rule mining. Research Journal of Recent Sciences, 4(3), 106-114.

Key findings (10 pts)

In 2-3 paragraphs, summarize the main findings in your source.

COPLINK is a data mining tool that helps crime analysts to identify patterns of crime in a given area. The tool performs functions such as data sharing, confining, and evaluating law-enforcement related information. Several approaches are integrated together to achieve the objective. For instance, the classification of data is achieved through a neural network, Naïve Bayes, decision tree, and Random Forest.

How do you know that this is a credible/scholarly source? (5 pts)

The six authors were students at Federal Urdu University of Arts, Sciences and Technology in Karachi, Pakistan when they were conducting this research. Their research was featured in the Research Journal of Recent Sciences, meaning it is peer-reviewed.

Reference 5

Citation in APA format (5 pts)

McClendon, L., and Meghanathan, N., (2015). Using machine learning algorithms to Analyze crime data. Machine Learning and Applications: An International Journal (MLAIJ) Vol.2, No.1, DOI: 10.5121/mlaij.2015.2101

Key findings (10 pts)

In 2-3 paragraphs, summarize the main findings in your source.

The Waikato Environment for Knowledge Analysis (WEKA), open source data mining software was used to conduct a study in the relation of violent crime patterns with actual crime data in different cities. McClendon and Meghanathan (2015) employed Additive Regression, Linear Regression, and Decision Trump algorithms with the same finite features on a given data set.

Basically, the researchers analyzed data in five steps namely; association, classification, clustering, forecasting, and visualization. Once data was fed to each of the three algorithms, it was analyzed by evaluating correlation coefficients, mean absolute errors, Root mean squared error, Relative absolute error, and Root Relative squared error. Linear Regression algorithm was found to be more effective and accurate in estimating crime data based on the input data set.

How do you know that this is a credible/scholarly source? (5 pts)

McClendon and Meghabathan conducted their research from Jackson State University in Mississippi, USA. Their work was thereafter published in the peer-reviewed journal Machine Learning and Applications: An International Journal.