Intelligent Systems for Analytics Assignment 3: MITS5509

Desklib provides solved assignments and past papers to help students succeed.

MITS5509 INTELLIGENT SYSTEMS FOR
ANALYTICS ASSIGNMENT 3
Student ID:
Student Name:

Table of Contents
Introduction................................................................................................................................2
Question 1..................................................................................................................................3
Decision Tree.........................................................................................................................3
Training Set................................................................................................................................3
Testing Set..................................................................................................................................5
Naive Bayes...............................................................................................................................7
Training Set............................................................................................................................7
Testing Set..............................................................................................................................9
Question 2................................................................................................................................11
Conclusion................................................................................................................................12
References................................................................................................................................13
List of Figures
Figure 1: Training Set Diagram.................................................................................................3
Figure 2: Output Screen.............................................................................................................4
Figure 3: Testing Set Diagram...................................................................................................5
Figure 4: Output Screen.............................................................................................................6
Figure 5: Training Set Diagram.................................................................................................7
Figure 6: Output Screen 1..........................................................................................................8
Figure 7: Output Screen 2..........................................................................................................8
Figure 8: Testing Set Diagram...................................................................................................9
Figure 9: Output Screen 1..........................................................................................................9
Figure 10: Output Screen 2......................................................................................................10
Figure 11: Dashboard...............................................................................................................11
1

Introduction
The prediction of the bankrupt problem is classified with the following datasets that are
provided which contains two ratios. It is computed with firm’s financial statements. For
evaluating or decreasing this problem, the classifiers will be used by training set and testing
set. Each of the data sets or data points are categorized by two categories (20 for category 0
and 20 for category 1).
2

Question 1
The organisation predicted the bankruptcy problem which can be resolved with the valid data
sets. The data sets are given for training sets and testing sets. It is included in the two ratios
which can be computed from organisation of the real world with financial statements. Data
sets of training sets contains almost 68 data values and it is categorized with the different
categories. As well as the testing sets also contains almost 68 data values. For reducing or
solving these problems, the different classifiers will be used in the training and testing sets.
Where 40 data values will be selected randomly, (Category 1 = 20 values, Category 0 = 20
values).
Decision Tree
It is the decision support process which uses the model or graph like a tree structure. In this
process, there are many possible consequences which includes the probability of the event
outcomes, utility and resource costs. It includes the control statement which helps to display
the proper algorithm (Anandarajan, et al., 2019).
Therefore, it is the group of nodes which are intended for creating decisions on the affiliation
of values to the classes or the numerical target values. Every node will represent the rule of
splitting for the single specific attribute.
Training Set
Figure 1: Training Set Diagram
In the above figure, it shows the design view of the training set. For making this design, there
are total 40 data sets included in the table. For evaluating the randomly selected data values,
3

the attributes will be selected by making connections with the training set module. After that,
the role is set for each and every attributes of the system. The data or the information of
training set is categorized into two categories such as 0 and 1. With the help of these
categories, the data is divided randomly.
The classifier named as decision tree used for evaluating the data or information. The
decision tree is connected with the split data function which is also connected with the apply
model function. The apply model is connected with the performance module or function
which helps to fulfil the connections of the whole process. The performance is the main
module or function used for getting the final result or values of each and every attribute
(Khan, et al., 2017).
Figure 2: Output Screen
Performance vector: It is the port which shows the criteria values of performance list which
is calculated with the attributes of label and the attributes of prediction of input Example Sets.
In the above screen, it shows the output screen of the training set module or table. Where it
results with the design of this module. It displays the performance vector of the design
module of training set. It shows the highest and the lowest values of the different attributes
which is displayed in the above image.
4

Testing Set
Figure 3: Testing Set Diagram
In the above figure, it shows the design view of the testing set. For building up the design, the
random 40 data values will be selected for evaluating the results of selected attributes. It has
the single connection between attributes and testing set or data. For the selected attributes, the
roles and responsibilities are selected for splitting the data in the different formats. The data
or information will be separated with the help of decision tree classifier. For further process,
the apply model is been used for making the valid connection with the performance. The
performance will help to getting the output by the design.
5

Figure 4: Output Screen
In the above figure, it shows the performance vector including the following criteria with the
highest and the lowest values or results. It almost covers 10+ criteria which have the both
highest and the lowest results. This result is segregated with the help of design made with the
help of data set or data values.
6

Naive Bayes
It is called as high-bias, or the classifier of low-variance. The main use of this classifier is to
build up the best model with the help of smaller data set or data values. It is very simple
classifier for use and the computationally inexpensive. It includes the spam detections, text
categorization, recommender system and sentiment analysis. It is the simple likelihood
classifier which is based on the Bayes’ theorem with strong assumptions of independence.
The main advantage or the benefit of the classifier Naive Bayes: It requires the small part of
the training data or information which estimates the variance and means of necessary
variables for every variable.
Training Set
Figure 5: Training Set Diagram
In the above figure, it shows the design view of the training set diagram with the help of the
classifier Naive Bayes. It has also the same process as compared with the decision tree
evaluation. In the performance of the classifier, it sets the error “Input set not having label
attribute” because of in the Set role function, the role has to be set as label attribute. The data
or information is separated with the help of set roles between Naive Bayes and apply model
(Upadhyay & Gautam, 2016).
7

Figure 6: Output Screen 1
Figure 7: Output Screen 2
In the above figure 6 and figure 7, shows the output screen of the data which is used in the
design view of training set for the Naive Bayes classifier. It is the performance vector of the
training set or values. It also shows the accuracy and the logistic loss of the organization with
the help of this classifier (Slater, et al., 2017).
8

Testing Set
Figure 8: Testing Set Diagram
In the above figure, it shows the design view of the testing set diagram with the help of the
classifier Naive Bayes. With the comparison of both the classifiers named Naive Bayes and
Decision Tree, basic difference occurs that the set role is initialized with the label attribute
but in the decision tree, the set role is initialized with regular attribute.
Figure 9: Output Screen 1
9

Figure 10: Output Screen 2
In the above figure 9 and figure 10, shows the performance vector results of the different
used attributes. It displays the output screen of the data which is used in the design view of
testing set for the Naive Bayes classifier.
10

Question 2
Figure 11: Dashboard
In the above figure, it shows the dashboard which is been developed by the raw data of
training set and testing set which is categorized by the WC, DC and Categories.
11