I graduated from Yildiz Technical University Statistics Department in 2018. I completed double major programme on Mathematics Engineering in 2019. I am currently working at Hepsiburada as a Customer Experience and Analytics Specialist. I generally do analysis with using SQL and Python, such as regression, churn prediction, satisfaction score prediction. I am interested in ML & DL Algorithms and Big Data areas. I want to enhance my abilities in this field, specially Supervised & Unsupervised Learning algorithms in the real world problems.
This video compares two different risk methods, Equal Risk Contribution and Hierarchical Risk Parity, to find the optimal portfolio, and explains their findings with machine learning methods. He uses the XGBoost method to compare metrics between two risk distribution methods. He also uses the SHAPforxgboost library to visualize its findings.
This post tries to find the customer segmentation with using the K-Means Algorithm. It also uses the Elbow Method and Average Sillhouette Method to find optimal cluster number. Finally, It decides 6 customer class is the most optimal class number. It got help from NbClust and factoextra libraries for visualizations.
This post examines how the sentiment towards presidential speeches has changed over time by analyzing emotions for a regular US speech between 1989 and 2017. A dictionary of positive and negative words was created and it was determined that the word was negative or positive based on these dictionaries for each word.
A movie recommendation algorithm was developed using the recommenderlab library on a dataset with 105339 scoring for 10329 films. For 2 users watching the same movie genre, a future movie that a user will watch will be offered to another. The logic of the suggestion system works in this way.