Description

Assignment 1

Dataset Choice:
Sentiment analysis using Bo and Pang’s movie review dataset with binary classification between positive and negative reviews.

Feature Design:
After getting the dataset some pre-processing was done on the dataset. All kinds of numbers, punctuations and special characters were removed. All the text was changed to lower case. After that I have selected bag-of-words as the feature design.
Bag-of-words (bow): A bag-of-words is a representation of text that describes the occurrence of words within a document. It involves two things: A vocabulary of known words and a measure of the presence of known words. In bag-of-words approach the number of occurrence and not sequence or order of words matters.
I have selected bow as the feature design as n-gram is mostly used for paragraph subjectivity classification and bow is mainly used for tweet or review classification. Since, the dataset of my choice was movie review classification so bag-of-words is justified.
A 5-fold cross validation was performed to select the best model.

CS5590 Foundations of Machine Learning Roll – EE19MTECH01008

Evaluation:
The accuracy for the different classifiers used is presented below in the form of tables.
Q1) K-NN classification.
p k 1 3 5
1 0.69 0.72 0.74
2 0.66 0.65 0.63
infinity 0.63 0.55 0.59

Q3) Kernelized SVM with RBF kernel.
GammaC 0.01 0.1 1 10 100
0.01 0.55 0.55 0.55 0.55 0.55
0.1 0.55 0.55 0.55 0.55 0.71
1 0.54 0.54 0.54 0.71 0.83
10 0.54 0.55 0.73 0.83 0.84
100 0.56 0.60 0.81 0.84 0.84

Q4) Regularized logistic regression.
C 0.001 0.01 0.1 1 10 100 1000
Accuracy 0.685 0.685 0.655 0.655 0.695 0.775 0.81

Q5) Gaussian based Bayes classifier.
Accuracy: 0.695

All codes are present in GitHub.