BEST SCORE ON KAGGLE SO FAR , EVEN BETTER THAN THE KAGGLE TEAM MEMBER WHO DID BEST SO FAR. The project is about diagnosing pneumonia from XRay images of lungs of a person using self laid convolutional neural network and tranfer learning via inceptionV3. The images were of size greater than 1000 pixels per dimension and the total dataset was tagged large and had a space of 1GB+ . My work includes self laid neural network which was repeatedly tuned for one of the best hyperparameters and used variety of utility function of keras like callbacks for learning rate and checkpointing. Could have augmented the image data for even better modelling but was short of RAM on kaggle kernel. Other metrics like precision , recall and f1 score using confusion matrix were taken off special care. The other part included a brief introduction of transfer learning via InceptionV3 and was tuned entirely rather than partially after loading the inceptionv3 weights for the maximum achieved accuracy on kaggle till date. This achieved even a higher precision than before.

A repository of resources for understanding the concepts of machine learning/deep learning. 

Corpus and a baseline neural network system for Named Entity Recognition in Hindi-English Code-Mixed social media text.

This clustering based anomaly detection project implements unsupervised clustering algorithms on the NSL-KDD and IDS 2017 datasets

Verifying suitability of dysphonia measurements for diagnosis of Parkinson’s Disease using multiple supervised learning algorithms.

Detecting drug-drug interaction (DDI) has become a vital part of public health safety. This project is an implementation of NLP based approach for such relation extraction between entities.

Automatic method for the recognition of hand gestures for the categorization of vowels and numbers in Colombian sign language based on Neural Networks (Perceptrons), Support Vector Machine and K-Nearest Neighbor for classifier /// Método automático para el reconocimiento de gestos de mano para la categorización de vocales y números en lenguaje de señas colombiano basado en redes neuronales (perceptrones), soporte de máquina vectorial y K-vecino más cercano para clasificador

📚 Use Google BERT on fake_or_real news dataset with best f1 score: 0.986

Extremely fast evaluation of the extrinsic clustering measures: various (mean) F1 measures and Omega Index (Fuzzy Adjusted Rand Index) for the multi-resolution clustering with overlaps/covers, standard NMI, clusters labeling

Report various statistics stemming from a confusion matrix in a tidy fashion. 🎯

Boundary F1 Score - Python Implementation

This is a simple python example to recreate classification metrics like F1 Score, Accuracy

With some projects to develop "TOOLs" for better Modeling

A hyperopt wrapper - simplifying hyperparameter tuning with Scikit-learn style estimators.

Script for calculating the optimal cut-off for max. F1-score or accuracy

This project aims to predict liver disease in Indian patients

Twitter Sentiment Analysis (Ml hackathon)

Classifying Forest Cover type

This is a web based elective course recommender system implemented with flask and Sklearn

Data Analysis in Spark to Identify Customer Churn for a fictional music service(like Spotify)

JS precision, recall and F1 calculation library

SVM for spamassassin statistical classifier plugin

Getting Intuition behind the implementation of Naive Bayes Classifier with SMS spam collection data.

Predict sales prices and practice feature engineering, RFs, and gradient boosting

GAN training a CNN.

Rank-based, collaborative filtering and matrix factorisation techniques for Recommendation Engine for IBM Watson Studio platform

Predict fraudulent credit card transactions using TensorFlow, Keras, K Neighbors, Decision Tree, SVM Regression and Logistic Regression classifiers .

Natural Language Processing (classification and machine translation) codes and analysis done for the year long practicum in Dublin City University (2019-20)

Custom metrics for LightGBM in R