This repository contains solutions to arbitrary data sets ranging from Machine Learning and Natural Language Processing to Computer Vision.
Color compression using K-means clustering:
Fragma Data Systems Assignment: Requires to build a binary Classification Machine Learning Model for the given marketing analytics data. Topics covered: Principal Component Analysis, Min-max Scaler, Machine learning classifiers, Grid Search Algorithm, Voting Classifier, Precision-Recall-Fscore based accuracy metric
Advenio Tecnosys: Requires an image segmentation task using the given image of retina. The main objective behind the task was to evaluate one's skills in thinking through the problem and the way or ways to solve it, particularly to see the approach towards solving the problem. Topics covered: Watershed algorithm, Color-space conversion using OpenCV
Named Entity Recognition Tagging using LSTM based neural network: A self-made tutorial explaining the use of Unidirectional and Bidirectional LSTM Models for NER tagging using wikigold.conll
data set. Topics covered: Unidirectional LSTM, Bidirectional LSTM, Keras Sequential Model, Confusion Matrix