alokkumary2j Goto Github PK

apostrophycorrector

This repository exhibits the power and the importance of domain knowledge(in this case, it's English grammar) in building good features and a good model. However, if you want to solve this problem today, my recommendation would be train a Bi-directional LSTM model - Primarily, because of the availability of ready made, free and massive amount of training data set for this problem(Pick up the content from any standard book, novel, news paper :-) ).

applied-text-mining

Each code snippet tries to tell a story from the Text Mining world.

codechefmay

codechefmaycookoff

Practice Solution for CodeChef May CookOff Challenge

fbflatposttypeclassification

Binary Classifier on Facebook Posts using Python NLTK: A simple Naive Bayes Classifier which obtains a given Flat/Apartment related Posts on Social Media websites (Facebook here) and classifies it whether it's made by Seller/Landlord/PG Owner or Buyer/Tenant/PG Seeker.

genericobjectdecoding

Demo code for Horikawa and Kamitani, "Generic decoding of seen and imagined objects using hierarchical visual features", https://www.nature.com/articles/ncomms15037.

hackerrankapril

hiddenmarkovmodels

Simple implementation of Hidden Markov Model for discrete outcomes/observations in Python. It contains implementation of 1. Forward algorithm 2. Viterbi Algorithm and 3. Forward/Backward i.e. Baum-Welch Algorithm.

hypertools

A python toolbox for gaining geometric insights into high-dimensional data

karza-recruitment-ml

At Karza Technologies, we are creating one of the most comprehensive and AI based digital due-diligence platform that helps financial institutes realise the possibilities of AI at each stage of decisioning viz on-boarding & KYC, due diligence, monitoring, skip tracing and recoveries. On top of thousands of publicly available data points, we are applying state-of-the-art algorithms/techniques in Fraud analytics, Semi-supervised Graph Algorithms, Computer Vision and NLP etc. to help financial institutes combat unconventional problems as diverse as capturing irregularities in Indian names & addresses, automate document worflows on virtually any kind of document and automated filtering of applications basis Risk Score. This Repository contains details pertaining to the Open-Paper Recruitment at Karza for the position of a Data Scientist.

linearalgebra

Visual Guide for Apparently Simple Linear Algebra Concepts using NUMPY

programmingassignment2

Repository for Programming Assignment 2 for R Programming on Coursera

psychopy

For running psychology and neuroscience experiments

pymvpa

MultiVariate Pattern Analysis in Python

r

Courseera

rapidautomatickpextractionrake

This repository contains my experiments with RAKE and its variants. RAKE is one of the most popular unsupervised approach for automatically extracting key-phrases/keywords from an unstructured data source like reviews, news, articles, documents etc.

sentiment-analysis-using-python-nltk

Sentiment Analysis Using Python NLTK. This Repository contains 3 models for Sentiment Analysis: 1. Naive Bayes 2. MaxEnt Model 3. SVM Model

spice---sequence-prediction-challenge

This repository holds an implementation of Discrete Hidden Markov Model which is trained and tuned to work for the sequence prediction challenge (SPiCe). Parameter tuning of this simple HMM implementation got me top 10 in the global ranking of the Sequence Prediction Challenge(SPiCe).

talking-data

Attempts on Kaggle Problem "Talking Data"

tensorsflowdeep

Primitive Examples/Experiments With Tensor Flow .