This repository holds some of my personal projects which I've done over the last few months. Currently, they're classified into below categories:
- Complex and original data science project containing: web scraping, data cleaning, exploratory analysis, model building and optimization: PEH_Classifier.
The aim of this project is to classify hair products into 3 categories according to PEH balance based on the ingredient list.
- Web Scraping and data cleaning project: Data_science_salary_predictions
The project contains web scrapped data from the glassdoor website with salary ranges. Dataset was cleaned and new features were extracted to prepare a dataset for future predictions. Min-max salary, company name, job state, and encoded skills: everything ready to start building ML model.
- Dimensionality reduction techniques comparison - Dimensionality_reduction
Used different classification models to achieve the most accurate classifier on high dimension dataset. Investigated how dimensionality reduction algorithms affect accuracy and learning time and make the dataset more understandable to the business.
- Classification problem with different classifiers and ensemble learning : Ensemble_learning_with_mushroom_dataset
Use different classifiers and their ensembles to recognize poisonous and edible mushrooms by their attributes. Use different accuracy metrics and classification reports to assess classifier accuracy. Examine feature importance for different algorithms.
Second notebook contains "Ensemble_learning_mushroom_class_pyforest" tests for importing ml&data science modules using pyforest library.
- Classification problem with different accuracy metrics and learning curves: Mushroom_classification
Build mushroom classifier with the highest precision and find the most indicative features.