scikit-learn-python

The practitioner's forecasting library

🔉 👦 👧Voice based gender recognition using Mel-frequency cepstrum coefficients (MFCC) and Gaussian mixture models (GMM)

A fast, robust library to check for offensive language in strings, dropdown replacement of "profanity-check".

🔉 👦 👧 👩 👨 Speaker identification using voice MFCCs and GMM

ML based Artificial Nose for spice identification using an array of 8 MOS gas sensors (MQ135, TGS series) with a PIC microcontroller. Implements Random Forest (100% accuracy) vs MLP comparison on 8135 sensor readings. IEEE conference paper, 2023.

Recognition of the images with artificial intelligence includes train and tests based on Python.

An end-to-end Machine Learning project that predicts student academic performance using data preprocessing, feature engineering, model training, and a Streamlit-based interactive web application.

Efficient sparse matrix implementation for various "Principal Component Analysis"

Octopi ML API is a Flask-based backend service hosting a machine learning model trained on over 500 software projects to predict development effort. It exposes RESTful endpoints for real-time effort estimation (in person-hours) and facilitates seamless integration with the Octopi frontend.

Serious AI forensic investigation workstation for tabular datasets: detects target leakage, structural corruption, and statistical anomalies with reproducible evidence and local AI.

skfeaturellm is a Python library that brings the power of Large Language Models (LLMs) to feature engineering for tabular data, wrapped in a familiar scikit-learn–style API.

The project scope is a weather forecasting model based on behavioral analysis of the last 33 hours (hour-by-hour forecast) with Random Forest Classifier. The program automatically saves and loads the last trained model for prediction.

Simple Python scripts that help automate and simplify tasks.

ML Python - Logistic Regression (GLM)

Crowd monitoring and cluster detection system with heatmap using Python & OpenCV

Real-time neural network observability platform for monitoring activations, gradients, information flow, and architecture health.

DMLLTDetectorPulseDiscriminator - A supervised machine learning approach for shape-sensitive detector pulse discrimination in lifetime spectroscopy applications

ML-powered phishing URL detection system — Random Forest, Gradient Boosting, SVM & Logistic Regression with a cyberpunk Flask web app

End-to-end machine learning classification pipeline on the Breast Cancer Wisconsin dataset featuring leakage-free preprocessing, GridSearchCV tuning, model evaluation, and SHAP explainability.

Codes for "Parkinson’s Disease Diagnosis: Effect of Autoencoders to Extract Features from Vocal Characteristics"