You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Utilizing machine learning models including logistic regression, random forest, gradient boosting, and neural networks to identify fraudulent credit card transactions. Dataset, consisting of PCA-transformed features and unbalanced classes, required precision-recall metrics for accurate evaluation. Developed in Python using TensorFlow and scikit.
Analyzed time-series data (Depressjon) to detect depression from patient activity recorded via clinical actigraphy watches. Utilized features such as time domain, statistical metrics, and LSTM-extracted attributes.
The goal is to eliminate manual work in identifying faulty wafers. Opening and handling suspected wafers disrupts the entire process. False negatives result in wasted time, manpower, and costs.
Analyzing Gujarat's education system with ML, we predict student dropout rates using demographic, economic, academic, and social data. Rigorous preprocessing, feature engineering, and model training aim to develop accurate dropout prediction models. Insights gained inform targeted interventions for dropout prevention.
This Python script visualizes the decision boundaries created by a linear Support Vector Classifier (SVC) on the Iris dataset. It utilizes scikit-learn for machine learning functionalities and matplotlib for plotting. The code loads the Iris dataset, trains a linear SVC on the first two features (sepal length and sepal width)
This is a binary classification problem. There are numerous factors that can contribute to the presence of heart disease. What is the most important factor causing heart disease? Can an accurate classifier be built to predict the presence of heart disease in patients? These are the questions we want to answer with this project.
This project focuses on building a fraud detection model for credit card transactions using a dataset containing transactions made by European cardholders in September 2013. We are working with a highly unbalanced dataset and the challenge lies in effectively detecting fraudulent transactions while minimizing false positives.