Sampling

In this assignment we had to first balance the unbalance data then divide the data into 5 samples using 5 different sampling techniques.After which we had to apply 5 different machine learning models on each sample and find out which ML model gave the highest accuracy.

Methods to Balance dataset

There are 2 methods to balance dataset :-

1. Undersampling - Sample the mojority class to the size of minority class

2. Oversampling - Sample the miority class to the size of moajority class

I have used Oversampling.

Size of Samples

I have used this Formula to calculate sample size

sample size = 384 Hence size for 5 samples = 384*5 = 1920

Sampling Techniques

I have used these 5 sampling techniques:-

• Random Sampling
• Systematic Sampling
• Cross-val Sampling
• Stratified Sampling
• Bootstrap Sampling

ML Models

Used the following 5 ML Classification Models :-

• LogisticRegression
• GaussianNB
• KNeighborsClassifier
• SGDClassifier
• SVC

Result

• Maximum value in the matrix: 98.18
• Row name at max value: KNeighborsClassifier
• Column name at max value: Cross-val