Hackaton of Lenta company at Skoltech.
Our implementation plan according to the downloaded notebooks:
- Encoding part. Here we decided to get rid of hash data and exchange them by encoding ones in every table.
- Clustering part. This part include forming new features for our data from RFMV approach and applying clustering KMeans algo to the customer data with new one.
- Preparing to the model. Before starting to train the model, it's not a simple way how data can be processed in order to be input for model. This part is devoted to this task.
- Training and testing model. In this part we train and test our model, get metrics, such as Precision@k and Recall@k metrices.
**5. Notebook MostPopularGoods.ipynb contain extra calculus for business model.
For each part we prepared well-organized notebooks. Check them out!