Fundamentals of Data Analytics

 

 

1. BACKGROUND
The purpose of this assignment is to implement the decision tree using
different split point evaluation measure and Naïve Bayes classification.
The performance metrics were evaluated for a machine learning model.
2. QUESTIONS
Question 1: (30points)
Given the following table construct a decision tree using a purity
threshold of 100%. Use information gain as the split point evaluation
measure. Present all the steps.
Next classify the points in terms of Risk:
a) (Age=27, Car= Vintage)- 15points
b) (Age=50, Car=Sports)- 15points
Point Age Car Risk
X1 25 Sports Low
X2 20 Vintage High
X3 25 Sports Low
X4 45 SUV High
X5 20 Sports High
X6 25 SUV High
Question 2: (30points)
Given the following table you need to predict the class label of a
tuple using Naïve Bayes Classification. Present all the probabilistic
calculations.
Customer Internet Service Contract Payment Method Churn
C1 DSL Monthly Cash No
C2 DSL Monthly Credit No
C3 Broadband Monthly Cash Yes
C4 Fiber Optics Monthly Cash Yes
C5 Fiber Optics Yearly Cash Yes
C6 Fiber Optics Yearly Credit No
C7 Broadband Yearly Credit Yes
C8 DSL Monthly Cash No
C9 DSL Yearly Cash Yes
C10 Fiber Optics Yearly Cash Yes
C11 DSL Yearly Credit Yes
C12 Broadband Monthly Credit Yes
C13 Broadband Yearly Cash Yes
C14 DSL Monthly Credit No
C15 DSL Monthly Credit ?
Next you need to predict churn or not churn for the following
customer:
C15= (DSL, Monthly, Credit)- 30points
Question 3: (20 points)
Consider the following confusion matrix. The numbers present if an
employee will leave or not from his/her organization.
Predicted Leave
Actual Leave No Yes Total
No 2326 15 2341
Yes 50 634 684
Total 2376 649 3025
Calculate the following metrics:
a) Accuracy
b) Sensitivity
c) Specificity
d) Precision
e) Recall
After the calculation, you need to discuss the results and provide
explanations regarding the performance of the Machine Learning
model selected.
Question 4: (20points)
Define the meaning of the following terms:
-Entropy-5pts
-Information Gain -5pts
-Gini Index-5pts
-Confusion Matrix-5pts

 

 

 

 

 

 

Sample Solution

 

exposition on facebook friendshipThe Internet these days assumes a critical job in individuals’ professions, connections, and different circles of life. Since it began to pick up ubiquity in the mid 1990s, it has transformed into a worldwide system, associating any person who can manage the cost of having a PC to the remainder of the world.

Bit by bit, administrations permitting to make new companions and to keep in contact with previously obtained companions began to show up, and today billions of individuals utilize different interpersonal organizations, of which the biggest is Facebook. These informal organizations despite everything stay a discussed wonder, just as the outcomes they lead to and the manner in which they have changed social orders. Furthermore, maybe, one of the weirdest (at any rate to me) marvels associated with them is Facebook companionship.

A Facebook kinship is adding an individual to your rundown of companions. As I would like to think, this is a sensible activity with individuals whom you care about, or whom you keep up a relationship with. Genuine companions, guardians, your beloved(s), partners with whom you spend time with after work, individuals whom you have warmed up to while voyaging, etc, should be available in any Facebook companion list.

Be that as it may, in reality, individuals include new individuals whom they have never met or known. Regardless of whether they have never traded words, they despite everything keep each other in their companion records. Or on the other hand another model is when associates who work in a similar office include one another, yet don’t speak with one another in Facebook or even in the workplace. Or on the other hand when irregular individuals who have (under certain conditions) conveyed for a few minutes, at that point add each other to their companion records. At times individuals even add different clients to their companion records for a demonstration of amount—maybe, it is intended to show how friendly they are. These, just as numerous other comparable cases, I don’t comprehend, and this is one reason why I have stopped utilizing informal organizations about a year back.

Companion records can be a wellspring of different aggravations. For instance, individuals some of the time are reluctant to erase such arbitrary “companions” from their rundowns, due to expecting an issues associated with this demonstration—managing someone’s feelings, for instance, or clarifying their reasons. Or maybe regularly, erasing individuals from Facebook companions should show the pace of dissatisfaction or outrage caused to a client by the erased individuals. Simultaneously, genuine correspondence frequently proceeds as though nothing spe

 

 

This question has been answered.

Get Answer