The project is aimed at applying Latent Dirichlet Allocation (LDA) with collapsed gibbs sampling to the Chip-Seq dataset across transcription factors(TF) to identify the interactions between TFs. Left-to-right and importance sampling evaluation techniques will be used to evaluate the model perplexity.
To evaluate the correctness of the inference techniques, we implement the exact inference for a toy dataset created from a multinomial distrbution. We then compare our results from exact inference to the left-to-right and importance sampling techniques.
Technologies Used: Python, Pandas, Numpy, Scipy
Implementation of SMO (Sequential minimal optimization) in Python.
Technologies Used: Python, OpenCV, Numpy
Predicted the price of the houses based on set of features using Linear Regression consisting of historic data from May 2014 to May 2015..
Technologies Used: Python, Pandas, Numpy
Created a program for Optical Character Recognition to predict numbers 3 and 5 (binary classification) using Perceptrons and Decision Trees.
Technologies Used: Python, Pandas, Numpy
Created a SVM model which can perform Sentiment analysis on the tweets and predict the correct sentiment of the text.
Technologies Used: Python, Pandas, Scikit-learn, NLTK
This course introduced me to the basics concepts in Machine Learning including regression, classification, and clustering. It also touched based on some of the concepts from Reinforcement Learning and Natural Language Processing.