DECISION TREE CLASSIFIER - HYPER PARAMETER TUNING - Binary Classification

Develop a technology for detecting mining sites using images from the optical satellite Sentinel-2. Specifically, it involves classifying images that contain mining sites and those that do not.

Implementing a model that can verify if two images belongs to same personality or not. Answer the question "Is this the claimed person?" It is a 1:1 matching problem i.e. given a face your task is to compare the candidate face to another and verify whether it is a match or not. My custom CNN model has achieved marvelous performance on the dataset.

A model for binary classification of credit card data as fraudulent or legitimate

This is heart disease prediction project that contains different methods such as FNN with Multiclass Classification, Binary Classification, Cross-Validation etc.

Used Machine Learning to create a cryptocurrency classification system for my investment banking client who is interested in offering cryptocurrencies for its customers.

Analysis of the credit history of a user and to predict if the user is a suitable candidate for a small ticket loan.

Duke University project: Data analysis in Excel (statistical analysis)

Histogram based classification and prediction of annual rainfall from Kerala dataset

Loan Default Prediction Model: A machine learning project that leverages historical lending data to create predictive models for assessing loan default risk, aiding financial institutions in making informed lending decisions.

Submission for Kaggle Competition "Real or Not? NLP with Disaster Tweet" (Top 10%) using SVM's, LSTM's, and BERT

A challenge on Predictive Analytics in Build With AI Hackathon 2021: A binary classifier to predict the entrepreneurial competency in University students

The code of the random forest project I shared on my Medium account

This nuget package is designed to help you easily identify and detect profanity or bad words within a given sentence or string. It works under a simple binary classifier that has been built and trained using ML.NET for accurate and efficient detection of inappropriate language.

Text Classification Problem : Wrote a module to classify Amazon-Product Reviews as favourable/unfavourable. Achieved accuracy of 78% and an F1 score of .81 using Logistic Regression on a test-train split of 20%, where total records were around 50000.

Social media fake accounts and spam accounts have become a huge problem these days. Some had spammed me twice on Instagram. Here I have used various Machine learning techniques to spot the fake/spam accounts

Create a machine learning model to determine the likelihood of a customer defaulting on a loan based on credit history, payment behavior, and account details.

Kyphosis disease prediction using Fully Connected Neural Networks (FCNNs) model and XGBoost model with GridSearchCV