Description

Foundations of Machine Learning (CS564 )
Department of CSE, IIT Patna

Marks:- 20 Marks

Instructions:
3. Be precise for your explanations in the report. Unnecessary verbosity will be penalized.
Prepare a Detailed report of the assignment.
4. Code should be done in Python or R.
5. You should zip all the required files and name the zip file as Group_no.zip, eg. Group_13.zip.
6. Upload your assignment (the zip file) in the following link: https://www.dropbox.com/request/ZJ0hsBb14A93EYx8wMi7

• The goal of this assignment is to experiment with feature extraction methods, linear methods for regression and logistic-regression.
____________________________________________________________________________ __
Brief Description of dataset:
Dataset Download Link: https://archive.ics.uci.edu/ml/machine-learning-databases/wine-quality/ Download winequality-red.csv

Dataset Name Wine Quality Dataset Characteristic Multivariate
Number of instance 1600 Attribute Characteristic Real
Number of Attribute 12

Attribute Description:

Attribute No Description
1 fixed acidity
2 volatile acidity
3 citric acid
4 residual sugar
5 chlorides
6 free sulfur dioxide
7 total sulfur dioxide
8 density
9 pH
10 sulphates
11 alcohol Output variable (based on sensory data)
12 quality (score between 0 and 10)

Instruction Regarding Dataset:
Apply 5-fold cross validation on the dataset.
(https://en.wikipedia.org/wiki/Cross-validation_(statistics))

Questions Linear Regression
1]Learn a linear classifier on the above dataset by using regression on alcohol variable
(feature no 11). Report the
a)Predicted alcohol content from learning(individual test data)
b)Report the average residual sum of squares (RSS) over these 5 folds.
(https://en.wikipedia.org/wiki/Residual_sum_of_squares)

Regularized Linear Regression
2] Use Ridge-regression on the above data on alcohol variable (feature no 11). Repeat the experiment for different values of λ(parameter). a) Report the residual error for each fold
b) Which value of λ gives the best fit?
( The value of lambda determines the importance of this penalty term. When lambda is zero, the result will be same as conventional regression; when the value of lambda is large, the coefficients will approach zero.)

Logistic Regression
3] Perform multi-class(0-10 classes) Logistic Regression on variable quality (feature no 12).
a) Report per-class precision, recall and f-measure for each fold.
b) Report the average per-class precision, recall and f-measure for 5 fold cross-validation.
c) Report the misclassification (provide confusion matrix) and also carry out the error analysis on few misclassified instances.