The data visualization created was to investigate the significant
differences and to demonstrate whether the measures are any
indication to produce techniques by recognizing a transitional
segment. j In the data visualization, the techniques provided included
the understanding of the descriptive statistics (summary) performed in
R programming language. This technique is conducted to assess the
construct validity of each input data respectively to the factor analysis
revealing a transitional phase segment and any significant correlation
between the various obtained functions. Furthermore, the descriptive
statistics function in R programming language provides an internal
consistency of reliability to determine an acceptable measure related
to the scores of a single factor. Below are three descriptive statistics
functions performed using R programming language to investigate this
methodology for further implication to define a visualization
technique (see tables below).
Table
. (Part 1 out of 3 tables presented). Descriptive statistics using R
programming language to perform functions of each input data
regarding minimum, the 1st quantile, the median, the mean, the 3rd
quantile, and the maximum values.
Table
. (Part 2 out of 3 tables presented). Descriptive statistics using R
programming language to perform functions of each input data
regarding minimum, the 1st quantile, the median, the mean, the 3rd
quantile, and the maximum values.
Table
. (Part 3 out of 3 tables presented). Descriptive statistics using R
programming language to perform functions of each input data
regarding minimum, the 1st quantile, the median, the mean, the 3rd
quantile, and the maximum values.
From the investigation of the descriptive statistics, this study offers
the ability to exercise the
StatExplore
feature in SAS Enterprise Miner. j
The described visualization includes the ability to examine the class
variation of the target factor and the associated variable for a defined
training output. To examine against the datasets, the function to
provides additional insight of the results showing reasonable trade-
offs between performance and accuracy of the factors. For example,
the class variation can be inherited to show that timely training is
achievable due to the dimensions of results. This approach was
explored to understand the training data of the class variable summary
statistics from the data role (train). The data role (train) can be
described using the variable name, role, number of levels, missing,
modes, and modes percentage (see figures below).
Figure
. Data diagram to describe insight of the class variation
according to percent variability and variable name of train data (e.g.,
approach, transition, and phase) using SAS Enterprise Miner.
Figure 8
. Data output to describe insight of the class variation
according to percent variability and variable name of train data (e.g.,
approach, transition, and phase) using SAS Enterprise Miner.
The figures above describe the insight using the
StatExplore
function
in SAS Enterprise Miner. j From the insight investigation, the mode and
mode percentages consider at least two modes to address the
functional (target) class with comparison. j The class variation also
includes the percent variability and the variable name of each train
data role according to the following variable names: a) approach (first
mode percentage is normal at 48.11%; second mode percentage is
steep at 38.81%) ; b) transition (first mode percentage is no at 95.80%;
second mode percentage is yes at 4.20%); and c) phase (first mode
percentage is final at 30.44%; second mode percentage is downwind
at 26.18%). This is vital to the ability to explore additional visualization
concepts with association to approach, transition, and phase. A factor
of (
No =
In the flight phase) determined and basic statistics generated
a calculation of 8031 subjects of the ID per a timestamp per
milliseconds. j As such, a factor of (Yes = the transitional phase)
determined and basic statistics generated a calculation of 3546
subject of the ID per a timestamp per milliseconds (see figure 9 below).
j j j
Figure
. Described
Transition Segment
(N= 2 mins per phase segment)
with a total score of 4 minutes as ID in R Programming Language.
The traffic pattern of an in the flight phase and the transitional phase
segment can be classified as a binary function. j The aim of this study is
too easy distribute the performed factors to be associated with the
application phase of each classified parameter. The factors included
the term
transition
and the examination to ensure the usefulness of
the application collected study data to monitor flight performances. j
This common baseline as a characteristic had accomplished the task to
export the data source to understanding the user interaction
according to the flight performance environments in a traffic pattern. j
The
No
factor is presented as
in the flight phase
(e.g., takeoff,
crosswind, downwind, base, final, and touchdown). j In comparison the
Yes,
this includes the
transitional phase
(i.e., per a 2 minute per phase
segment to total 4 minutes, which this study accounts for a transition
as described in the figures below
Figure
. Described
Phase
of a Traffic Pattern associated with the
timestamp to preformed as ID in R Programming Language.
Figure
. Described
Approach
of a Traffic Pattern associated with the
timestamp to preformed as ID in R Programming Language.
Based on the train variable names, the data measures with respect to
the underlying distribution of each input values introduce methods for
characterization and comparability. j These properties are proposed as
a statistical method to visualize and interpret the graphical inference
tools according to the baseline characteristics. The added input can be
export from the used data capture system of the function label
transition
that includes the implementation of a factor analysis
individually for each data input. The function
transition
has been
defined as the transitional factors regarding the phases of flight (in a
traffic pattern) according to performance and functionality. The
action to further investigate the described insights include the
exploration of a plot matrix and the class association (e.g., no and yes
of the transitional traffic phase segment) to each variable input. j The
figures below are the visualization using Weka Explorer according to
the transition and the selected instance of the data representation (i.e.,
approach, transition, and phase) (see figures below) .
Figure
. Phase (variable name) described insight using a plot matrix
visualization according to the Phase (x: variable name) and Transition
(y: factor) with the class color (e.g., no and yes) in Weka Explorer.
Figure
. Approach (variable name) described insight using a plot matrix
visualization according to the Approach (x: variable name) and
Transition (y: factor) with the class color (e.g., no and yes) in Weka
Explorer.
Figure
. Transition segment (variable name) described insight using a
plot matrix visualization according to the Transition (x: variable name)
and Transition (y: factor) with the class color (e.g., no and yes) in Weka
Explorer.
This discovery had altered the scope to focus the data analysis
according to three elements (e.g., approach, transition, and phase) to
understanding the transition segment of a traffic pattern. The
correlation among the traffic pattern transition segment considers the
dynamics expressed in short-term of dependences and the target
variables. These proposed target variables can be explored to predict
mutual information and the performance selected of predicted results.
The results provided can show the proposed methods to achieve the
necessary action as expressed of each transition segment significant
to the traffic pattern in flight operation near airports. From the
evaluation, the three elements allowed for the ability to address the
characteristic behavioral patterns and the scalability of the dataset
contains j The Naïve Bayes classifier was consider understanding the
pattern mining techniques according to the criteria evaluation on
training set (see table below).
Table
. Evaluation on training setoff the Summary with Detailed
Accuracy by Class using a Confusion Matrix in Weka Explorer.
j j j The features are detailed with accuracy by class as this discovery
allowed for the confusion matrix to be classified as
no and yes
. The
research work will include the proposed weighted average as the
selection algorithm to balance the various condition for predictive
modeling. j This proposed technique can improve the overall
performance in terms of accuracy for model development and
evaluation of the training set. j In comparison, the unweighted versus
weighted model is significant to capture the effects to create a fair and
balanced classification. Without consideration, this altered condition
can lead to an unbalanced cluster of classified instances. Therefore,
the benefit allows for a heuristic consideration to analyze behaviors of
the target class using the statistical features with a level of accuracy
proposed. This will effective in understanding the attributes and how
the future evaluation on training set can be influenced (see figure
below of all the attributes plots).
Figure
. All the attributes plot for evaluation in Weka Explorer.
