My Latest Projects
I have worked on multiple projects projecting the latest technologies and algorithms. Shown below are some of
those projects. For the source code please check out my page at GitHub!
Image Classification
Implemented a Convolution Neural Network (CNN) model with Keras to predict the classification of types of images using a CIFAR-10 dataset and procured a model accuracy of 75%.
PREDICT INJURY SEVERITY
Applied a Feed Forward Neural Network made from scratch using only Pandas libraries in python to a FARS dataset to examine relationships between driver injury severity and roadway condition to obtain an accuracy of 46% for the multi-class classification and 60% for the binary classification
RANDOM MATRIX SPARSIFICATION
Analyzed and compared the performance of K-Means on the data matrix using 3 separate sets of data.Used K-Means++ initialization method algorithm to obtain clusters which along with pre-processing steps which led to an increase in performance by 60% .
Multi-label Categorization of construction projects
Successfully extracted and labelled information such as construction, design, operation and maintenance from construction contracts and project requirement documents using supervised learning in python.
Image Segmentation
Trained a machine learning model to segment an image into multiple objects built using a dataset of ~12K images and 20 classifications to obtain an accuracy of 84.5 %.
Movie Recommendation System
Build a content based movie recommendation system and make an API using django and deploy it on the cloud.