Towards Financially Inclusive Credit Products Through Financial Time Series Clustering

By Tristan Bester, Benjamin Rosman

Rating

1361
Battle Count: 15

Relevance

7/10
While focused on customer segmentation, the time series clustering techniques could be adapted for market segmentation or trading strategy development

Implementation Complexity

6/10
Requires deep learning expertise and careful tuning of learning rates, but builds on existing architectures and techniques

Reproducibility

4/5
Detailed methodology and component configurations provided, but exact hyperparameters not fully specified

About this paper

Methodology: Financial Transaction History Clustering (FTHC). Problem types: Clustering, Time Series Analysis, Dimensionality Reduction.

The interactive Everscope explorer (charts, battles, favorites) loads below.