A dataset OD_dataset.csv with the aggregated transactional data for merchants (AKA entities) belonging to the "Fast Food Restaurants" category is given. The aggregation is done by grouping transactional data by the [entity_id] column. You can make an assumption that there are approx. 1% of outliers. This dataset contains 33 columns.
- Perform exploratory analysis of the dataset. Find outliers / anomalies using any unsupervised learning algorithm(s) / ensembles (at least two methods). One of the method should make use of deep learning.
Used techniques:
- PAM clustering
- Autoencoder
- Visualize the data and detected outliers using at least two different dimensionality reduction techniques.
Used techniques:
- PCA
- t-SNE
- Visualizations of results done in Plotly