Aman Gupta

House Prices: Advanced Regression Techniques

Applied various regression algorithms to accurately predict the price of houses. Problem Statement

List of algorithms applied :

Linear regression using ‘lbfgs’ optimization algorithm
Support Vector Regression(SVR) using linear kernel
Random Forest Regressor
Gradient Boosting Regressor

The best accuracy was obtained using Gradient Boosting Regression and the obtained Root Mean Squared Logarithmic Error(RMSLE) over the test cases was 0.14065 Click Here For Code

Titanic: Machine Learning from Disaster

Applied various classification algorithms to accurately predict whether a person with given features survived the Titanic boat. Problem Statement

List of algorithms applied :

Logistic regression using ‘lbfgs’ optimization algorithm
Random Forest Classifier
Gradient Boosting Classifier

The best accuracy was obtained using Random Forest Classifier and the obtained Categorisation Accuracy over the test cases was 77.511% Click Here For Code

Binary search Tree and Image Editor

Implemented Binary search Tree and build an Image editor Problem Statement

Part 1: Binary Search Tree

Implemented a Binary search tree having the following in built function :

findMin : Returns the minimun element of the tree
insert : Inserts an element into the tree
deletemin : Deletes the minimum element of the Tree
k-sum : Takes a list an input and returns the sum of the smallest k elements of the list . This is done using a binary search tree

Part 2 : Image Editor

The image editor can perform the following function on a pgm(Grey Scale) image :

Averaging Filter : This function is used to blur the image by replacing each pixel by average of pixels surrounding that cell.
Edge Detection : This function detects the sudden change of pixel in an image so as to detect the Boundary of an object in the image . This is done by wrapping the image and the normalizing the gradient image .
Path of least energy: In this function we find the path of least energy in the image so that we can crop the unnecessary parts of photo and focusing on main parts(Seam Carving) . This is done by Dynamic programming algorithm , following the top-bottom and bottom-up approach

This assignment was done under Prof. Prem Kalra
Click Here For Code