natalliakarnilava Goto Github PK

accident_risk

The purpose is to understand whether it is possible to predict the risk of an accident based on the historical data of one of the regions. The project implemented an analysis of the main factors of an accident using SQL, LightGBM, Seaborn, Matplotlib, etc.

hotel_chain_client_churn_interim_project

mllib

In the project, a linear regression model is trained on California housing data in 1990. The goal is to predict the median value of a house in a residential area. The metrics RMSE, MAE and R2 are used to assess the quality of the model. As well as pyspark/mllib methods were used.

neural_network_building

prediction of an absolute temperature on the surface of a star using neural networks

personal_data_protection

The purpose - to protect the personal data of the insurance company's clients and to develop a data transformation method that makes it difficult to recover personal information. R2 metrics of Linear Regression's do not worsen during the transformation.

sentiment_analysis

Purpose is to create a tool, which will be able to classify comments as to whether they are toxic or not. Used SpaCy, NLTK, Logistic Regression and GridSearchCV to achieve F1-score higher than 0,75 on the test sample.

sp_pr

supervised_learning

time_series

Taxi demand forecasting for the next hour , given historical data, as well as calendar signs, previous values, and a moving average.