Caner Burç BAŞKAYA's repositories
Emotion-Recognition-Out-Of-Facial-Expression-With-Deep-Learning
This finished product is able to analyze, evaluate and recognize the 7 fundamental emotions. The product can process live video streams as well as recorded videos utilizing convolutional neural networks (CNN) with Keras and is deployed using Flask.
European-Credit-Card-Fraud-Detection
This projects aims to detect the fraudulent transactions and to avoid them before happening to prohibit loss in millions of $. The prediction model succeeds a 97.3 precision, 92.7 recall rate alongside 93.6% F1 score with beta 2.0 for validation set.
Pharmaceutical-Products-Sales-Forcasting
This project aims to predict the sales volume in units of 5 pharmaceutical products for 4 months including Covid-19 period. The work fulfilled the best prediction scores in a competition with MAPE scores ranging from 3% to 11% for 4 products. The company & product names are private due to confidentiality.
Sophia-Chat-With-PDF
Open AI GPT Q&A ChatBot. Developed using LangChain, OpenAI - GPT3.5 Turbo, Facebook FAISS, Streamlit, Azure
Animal-Kingdom
A cross platform mobile application written in Flutter\Dart in which the app guesses the animals users picked in their minds based on a decision tree. The game also has the ability to improve itself.
Anomaly-Detection-For-Sustainable-Production-in-Oil-Wells
An anomaly detection model to avoid unpredictable and displeasing event by countermeasures before it happens | note: unfinished
Churn-Prediction-for-U.S.-Telecom-Data
It is focused on customer churn utilizing the personal info and various user statistics on a telecom user database from U.S. The performance of the final model is 98.8% accuracy and 96.4% f-beta score so far.
Crisis-Management-Performance-On-Volswagen-Emmision-Scandal-In-The-Light-Of-NLP-and-Sentiment-Analys
My-Online-Tutor
An Online Commercial Tutoring Platform in which users can get prepared to university admission exam with the help of professional lecturers.
Online-Marketing-Purchase-Prediction
This work predicts the buyers and non-buyers on an online shopping platform at a 92.4% accuracy, 89.8% precision and 95.3% recall performance.
IBM-Fundamentals-of-Scalable-Data-Science---Apache-Spark---pySpark
IBM Fundamentals of Scalable Data Science on Coursera
Libros-Para-Leer
It is an online mobile application in which users purchase and read more than 1000 books in Spanish comfortably. You can browse it on store via the link as well.
Soccer-Players-National-Player-Prediciton
This projects aims to predict if a soccer player has participated in his national team or not. The prediction model succeeds a 95.1% f-beta score for training and 87.4% for validation set.
World-Flag-Quiz
An Mobile Android Game, in which the players guess the flags of world countries using many lifelines and power-ups, in addition to purchase packages and study the flags with built-in tutorial centre.