FDHelper

Assist Unsupervised Fraud Detection Experts with Interactive Feature Selection and Evaluation

 

 

CollisionDistanceVisualization.png

 

Image:The interfaces of FDHelper. The model configuration panel (a), multi-layer visualization map (b), selected group inspection (c), brush and link selected users for details (e) and adjust the feature sets selection panel (d).

 

 

OBJECTIVE

1. Design and implement an end-to-end interactive visualization system, FDHelper

2. Identify a workflow based on experience from both fraud detection algorithm experts and domain experts.

3. Use a multi-granularity three-layer visualization map embedding an entropy-based distance metric ColDis, analysts can interactively select different feature sets, refine fraud detection algorithms, tune parameters and evaluate the detection result in near real-time.

 

PUBLICATION

·        Jiao Sun, Yin Li, Charley Chen, Jiahe Lee, Xin Liu, Zhongping Zhang, Ling Huang, Lei Shi, Wei Xu,“FDHelper: Assist Unsupervised Fraud Detection Experts with Interactive Feature Selection and Evaluation”, CHI, 2020. [paper] [code]