Credit Scoring Model
Implementation of a probability evaluation system of default credits on the base of bank own data and the banking credit history over a given time.
Several types of classifiers have been implemented, including SVM and kNN, neural network, logistic regression and some others, as well as algorithms to select optimal parameters for the evaluation. The method of cross-validation was chosen to find the optimal parameters for each of them by taking risk assessment and expected profit into account.
The results were visualized.
The final solution was designed in the form of a program for a computer-aided decision-making.
- Data Classification
- Prediction Modeling
- Feature Selection
Virtual try-on tool for makeup products
The system consists of a face detection and segmentation model and an algorithm that allows recoloring objects without losing their original texture.
Online sign language interpreter
AI algorithm that converts video of a person using sign language into a text transcript
Workout helper app
Mobile app for the estimation of proper body positions during the workout.