FarzadNekouee
Computer Vision Engineer | Expert in Machine Learning, Deep Learning & Data Science
Isfahan, Isfahan Province, Iran
Pinned Repositories
Flight-EDA-to-Preprocessing
An extensive exploration and preprocessing of Flight data. The project encompasses detailed EDA (Univariate, Bivariate, and Multivariate analysis), along with comprehensive data preprocessing techniques - missing value treatment, outlier management, categorical encoding, feature scaling, and skewness transformation.
Gold-Price-Prediction-LSTM
Harnessing LSTM for gold price prediction from 2013-2022. Trained on 9 years of data, the model predicts prices for 2022. The repo contains detailed explorations, visualizations, and insights, achieving high accuracy evaluated by RMSE.
Heart_Disease_Prediction
In this repo, we analyze a dataset of heart patient metrics to build a model identifying heart disease risks. We prioritize high recall for comprehensive detection through EDA, preprocessing, and model building. Explore our approach and findings!
Hotel_Booking_Cancellation_Prediction
Predictive modeling of hotel booking cancellations. Handling complex data preprocessing, feature engineering, and noise. Evaluating models for high F1-score and analyzing key features for interpretability.
Keras-CIFAR10-CNN-Model
An exploration of CIFAR-10 image classification using a CNN model built with Keras. This repository includes training, evaluation, and external image testing to understand model generalization.
Retail_Customer_Segmentation_Recommendation_System
Analyzing and transforming a UK-based retail dataset (2010-2011) into a customer-centric format for customer segmentation using K-means clustering. Implementing a personalized recommendation system to enhance marketing strategies and boost sales.
Telegram-IMDb-Bot
Telegram IMDb Bot: Python-based bot for Telegram. Features include movie search via IMDb API, watchlist management, and movie details retrieval. Check the demo video to see it in action.
Traffic-Violation-Detection
An urban traffic violation detection system using classical image processing techniques. Features include real-time traffic light recognition, adaptive night-time stop line detection, robust license plate extraction, PyTesseract OCR for text recognition, dynamic penalized plate display, and MySQL logging.
YOLOv8_Pothole_Segmentation_Road_Damage_Assessment
Automated pothole detection and road damage assessment using YOLOv8-seg for real-time analysis.
YOLOv8_Traffic_Density_Estimation
A YOLOv8-based project for real-time traffic density estimation. It employs fine-tuned vehicle detection models to analyze and count vehicles per frame, aiding urban traffic management and planning. The repository includes model training, traffic intensity analysis, and deployment strategies.
FarzadNekouee's Repositories
FarzadNekouee/Retail_Customer_Segmentation_Recommendation_System
Analyzing and transforming a UK-based retail dataset (2010-2011) into a customer-centric format for customer segmentation using K-means clustering. Implementing a personalized recommendation system to enhance marketing strategies and boost sales.
FarzadNekouee/YOLOv8_Pothole_Segmentation_Road_Damage_Assessment
Automated pothole detection and road damage assessment using YOLOv8-seg for real-time analysis.
FarzadNekouee/Telegram-IMDb-Bot
Telegram IMDb Bot: Python-based bot for Telegram. Features include movie search via IMDb API, watchlist management, and movie details retrieval. Check the demo video to see it in action.
FarzadNekouee/Traffic-Violation-Detection
An urban traffic violation detection system using classical image processing techniques. Features include real-time traffic light recognition, adaptive night-time stop line detection, robust license plate extraction, PyTesseract OCR for text recognition, dynamic penalized plate display, and MySQL logging.
FarzadNekouee/YOLOv8_Traffic_Density_Estimation
A YOLOv8-based project for real-time traffic density estimation. It employs fine-tuned vehicle detection models to analyze and count vehicles per frame, aiding urban traffic management and planning. The repository includes model training, traffic intensity analysis, and deployment strategies.
FarzadNekouee/Heart_Disease_Prediction
In this repo, we analyze a dataset of heart patient metrics to build a model identifying heart disease risks. We prioritize high recall for comprehensive detection through EDA, preprocessing, and model building. Explore our approach and findings!
FarzadNekouee/Gold-Price-Prediction-LSTM
Harnessing LSTM for gold price prediction from 2013-2022. Trained on 9 years of data, the model predicts prices for 2022. The repo contains detailed explorations, visualizations, and insights, achieving high accuracy evaluated by RMSE.
FarzadNekouee/Keras-CIFAR10-CNN-Model
An exploration of CIFAR-10 image classification using a CNN model built with Keras. This repository includes training, evaluation, and external image testing to understand model generalization.
FarzadNekouee/Flight-EDA-to-Preprocessing
An extensive exploration and preprocessing of Flight data. The project encompasses detailed EDA (Univariate, Bivariate, and Multivariate analysis), along with comprehensive data preprocessing techniques - missing value treatment, outlier management, categorical encoding, feature scaling, and skewness transformation.
FarzadNekouee/Facial_Analysis_with_Mask_Detection
AI-powered tool for Facial Attribute Analysis. It accurately predicts age, gender, race, and emotions for unmasked faces. Enhanced to detect masked faces, ensuring reliable analysis in real-world scenarios. Utilizes DeepFace for in-depth insights.
FarzadNekouee/Garbage_Classification_ResNet50_Scratch_to_Transfer-Learning
In this project, we harness ResNet50 for garbage classification. Using a dataset pivotal for automating waste segregation, we address challenges like class imbalance with data augmentation and tailored metrics. The journey involves analyzing the dataset, confronting imbalances, and developing models from scratch and via transfer learning.
FarzadNekouee/Hotel_Booking_Cancellation_Prediction
Predictive modeling of hotel booking cancellations. Handling complex data preprocessing, feature engineering, and noise. Evaluating models for high F1-score and analyzing key features for interpretability.
FarzadNekouee/XGBoost-CatBoost-Employee-Resignation
The project aims to predict employee resignations using XGBoost and CatBoost regressors. It involves extensive EDA, data preprocessing, model building, hyperparameter tuning, and model evaluation. The objective is to help organizations take proactive steps to retain valuable employees and maintain stability.
FarzadNekouee/DCGAN-Photorealistic-Face-Generator
This project leverages DCGANs and the CelebA dataset to create photorealistic, synthetic human faces. The DCGAN model crafts new faces that, although similar to the celebrities in the training set, are completely fabricated. The goal is to produce lifelike 64x64 pixel faces.
FarzadNekouee/Imbalanced_Bank_Loan_Modeling
Classification project to pinpoint potential loan customers from an imbalanced dataset. Emphasizes on penalized and tree-based algorithms, optimizing for both recall and precision to enhance campaign efficacy and conversion rates.
FarzadNekouee/Regression-Assumptions-Regularization
Dive into polynomial regression and its assumptions in car price prediction. This repo explores Linear, Polynomial, Ridge, Lasso, and Elastic-Net models on the CarDekho dataset for accurate estimates. A deep dive into regression and regularization.
FarzadNekouee/Enhanced_MRI_Tumor_Classification_Web_App
Reliable MRI brain tumor detection for various types including pituitary, meningioma, glioma. Focused on high accuracy in diverse image qualities. Features a Streamlit web app for practical deployment.
FarzadNekouee/Mobile_Price_Range_Classifier
Launching a mobile company to rival giants like Apple & Samsung, we leverage sales data to discern price ranges for mobiles based on features via ML models like SVM, DT, & RF. Not aiming for exact prices, but a strategic price bracket.
FarzadNekouee/WeatherProof_PetDetection_App
This project presents a weather-proof model for accurate detection of dogs and cats, streamlined for desktop application. Utilizing renowned models, transfer learning, data augmentation, and weather effects, we've achieved a robust solution ready for real-world deployment.