Hey

Hey there it's Erdogan. I am interested in AI, Machine Learning, Deep Learning.

Some of my Notes:

• Machine Learning (Andrew Ng.)
• Deep Learning Specialization (Andrew Ng.)
• Tensorflow in Practice (Laurence Moroney)
• FastaiV3-Part1 (Jeremy Howard)
• FastaiV4-Part1 (Jeremy Howard)
• Basic Python Machine Learning
• Undergraduate Thesis - Exoskeleton
• School Projects
• Angular Notes
• Web Scraping
• SQL Basics
• Python Basics
• Data Structures and Algorithms
• Git Notes

Some of my Projects:

Movie Recommendation App

• Machine learning based movie recommendation app. Users can combine different movies and get suggestions for the provided combination. I trained a collaborative filtering model and obtain embedding vectors for about 60k movies. Recommendations relies on these embedding vectors. Designed a simple front-end with angular and backend with flask.
• active website
• github page

Joke Recommender

• Again I trained a collaborative filtering model on some jokes. This recommendation system works a little different than the above. I ask users to rate some jokes first, and based on these jokes, system creates a taste profile for the user (embedding vector of the most similar user based on the ratings). The recommendations is done based on this taste profile.
• Binder App:
• github page

Collaborative Filtering Notebook

• notebook

Tabular Models Notebook

• notebook

NLP Notebook

• notebook

"Beauty" Detector

• It is a computer vision application. Basically the app tries to detect "beauty". This project is done only for educational purposes and fun.
• I collected data using google search. Old version does a binary classification and assign a probability for "beauty" based on the given image.
• So it was basically a binary image classificiation model.
• Old Version Model Training
• Later I improved this app, added it some other features. Two different model works on it. One is a multiclass image classifier (based on gender and "beauty" level). Another is an image regressor (age prediction).
• Binder App:
• github page

Price Predictions

• Sales Prediction Notebook
• Weekly Price Prediction Test

Survey App (Angular)

• It is a simple survey app that I've created for a friend that allows users to create surveys and share it with others. I used Angular. It uses Firebase to store the data
• deployed app page
• github page

Some Websites

• erdoganyildiz.com
• old blog page mvc github

Crypto Price Compare Bots

• Some web crawlers that compares crypto prices on different crypto markets and send notification e-mails.