Pinned Repositories
Automatic-Speech-recognition-ASR-
Feature vector for Automatic Speech recognition :Mel Frequency Cepstral Coefficient (MFCC)
Decision-Tree-in-Python-for-Continuous-Attributes
This code constructs a Decision Tree for a dataset with continuous Attributes. Each training instance has 16 numeric attributes (features) and a classification label, all separated by commas. In deciding which attribute to test at any point, the information gain metric is used. The node test threshold for each potential attribute is set using this same metric i.e. at each point, all the values that exist for a particular attribute in the remaining instances are ordered, and threshold values that are (half way) between successive attribute values are used to find the Information Gain. The threshold value that gives the highest information gain is used. The same attribute can be tested again later in the tree (with a different threshold).
ensemble-of-sarima-random-forests-and-gradient-boosting-trees
In this Project I use the Kaggle Bike sharing dataset to predict the sales of bike given a Multivariate Time series. I model the multivariate data using ensemble of Random Forests and Gradient Boosted trees. After that the residuals of the model are fit with an ARMA/ARIMA/SARIMA model and later forecasted. The residuals are later added back to the predicted values
genetic-algorithm-for-cnn
This project tunes a Convolutional Neural Network using a genetic algorithm for Image Classification.
graph-based-semi-supervised-learning
This project explores the different techniques (both scalable and non scalable) for Graph based semi supervised learning. Recent techniques such as ITML and LMNN along with a few others are empirically evaluated on the 20 newsgroups dataset.
Hidden-Markov-Model
This Code Implements the Hidden Markov Model (Monitoring and the Viterbi Algorithm) in Python on a Time series Data.
Machine-learning-algorithms
SMO algorithim for training support vector machines,Independent component analysis,Principal Component analysis
Tic-Tac-Toe-Using-Alpha-Beta-Minimax-Search
This code demonstrates the use of Alpha Beta Pruning for Game playing. Since, Tic Tac Toe has a depth of 9 , I use a heuristic function that evaluates the Board State after searching through a depth of 3. The heuristic function calculates the expected score of winning for the PC given the board state.
tsp-using-simulated-annealing-c-
This code solves the Travelling Salesman Problem using simulated annealing in C++.
Viterbi-algorithm
The Viterbi algorithm is tagging algorithm based on TRIGRAM HIDDEN MARKOV MODELS (TRIGRAM HMMS)
deerishi's Repositories
deerishi/graph-based-semi-supervised-learning
This project explores the different techniques (both scalable and non scalable) for Graph based semi supervised learning. Recent techniques such as ITML and LMNN along with a few others are empirically evaluated on the 20 newsgroups dataset.
deerishi/Tic-Tac-Toe-Using-Alpha-Beta-Minimax-Search
This code demonstrates the use of Alpha Beta Pruning for Game playing. Since, Tic Tac Toe has a depth of 9 , I use a heuristic function that evaluates the Board State after searching through a depth of 3. The heuristic function calculates the expected score of winning for the PC given the board state.
deerishi/ensemble-of-sarima-random-forests-and-gradient-boosting-trees
In this Project I use the Kaggle Bike sharing dataset to predict the sales of bike given a Multivariate Time series. I model the multivariate data using ensemble of Random Forests and Gradient Boosted trees. After that the residuals of the model are fit with an ARMA/ARIMA/SARIMA model and later forecasted. The residuals are later added back to the predicted values
deerishi/genetic-algorithm-for-cnn
This project tunes a Convolutional Neural Network using a genetic algorithm for Image Classification.
deerishi/Decision-Tree-in-Python-for-Continuous-Attributes
This code constructs a Decision Tree for a dataset with continuous Attributes. Each training instance has 16 numeric attributes (features) and a classification label, all separated by commas. In deciding which attribute to test at any point, the information gain metric is used. The node test threshold for each potential attribute is set using this same metric i.e. at each point, all the values that exist for a particular attribute in the remaining instances are ordered, and threshold values that are (half way) between successive attribute values are used to find the Information Gain. The threshold value that gives the highest information gain is used. The same attribute can be tested again later in the tree (with a different threshold).
deerishi/Hidden-Markov-Model
This Code Implements the Hidden Markov Model (Monitoring and the Viterbi Algorithm) in Python on a Time series Data.
deerishi/coursework
deerishi/Policy-Search-in-a-Markov-Decision-Process
This code evaluates an optimal policy in a Markov Decision Process. We use a 3x3 Grid World with the Goal State at 3,3 with a reward of 10 and the rest of the non terminal states with a reward of -1.
deerishi/tsp-astar
This code solves the Travelling Salesman Problem using Astar Search. Minimum Spanning Tree Heuristic was used to estimate the remaining distance from one city to the last.
deerishi/ADOC
deerishi/assignment5ML
deerishi/Bayes-Net-Structure-Prediction
Learning how to predict a Bayes Net Structure of a Dataset
deerishi/Bernoulli-Document-Model_Based-Naive-Bayes-SMS-Spam-Classification
This code is for Naive Bayes Spam Classification on the SMS Spam Collection Data Set from the UCI Machine Learning Repository.
deerishi/codechef
deerishi/finalPcOS161
deerishi/learning-nodejs-10-projects
Following the "Learning NodeJS By Building 10 Projects" on Udemy
deerishi/Logistic-Regression-Convergence-Analysis
This code implements Logistic Regression using Newton's Method in Python.
deerishi/MLPTrain
deerishi/Non-Linear-Kernelized-Regression
This code implements Non Linear Kernelized Regression using a Gaussian Kernel on a dataset
deerishi/old_pc_OS161
deerishi/PacMan
Multi Agent Pacman Berkley 188
deerishi/server_version_OS161
deerishi/setjmp-longjmp-ucontext-snippets
Implementing coroutines, channels, message passing, etc.
deerishi/stat929
deerishi/sudoku-as-graph-coloring-and-constraint-satisfaction-problem
deerishi/t-SNE
This code demonstrates the how to use t-SNE from Scikit -Learn's implementation
deerishi/Telstra-Network-Disruptions
deerishi/tensorflow
Computation using data flow graphs for scalable machine learning
deerishi/Time-Series-Project
deerishi/variable-elimination
Variable Elimination for Bayes Net