image77
image78
  • Home
    • How Does it Work?
    • Schedule a Demo
    • COVID-19
    • Finance
    • Other
  • Resources
    • About Us
    • Careers
    • Investors
    • Home
    • Product
      • How Does it Work?
      • Schedule a Demo
    • Applications
      • COVID-19
      • Finance
      • Other
    • Resources
    • Company
      • About Us
      • Careers
      • Investors
  • Home
  • Resources

Akai Kaeru's Explainable AI for Fintech Analysis

We took a data set including stocks in the S&P 500 to analyze monthly returns from January to February 2016. The goal was to identify the key drivers of return during this time period. The data included over 25 features describing different aspects of the stocks (e.g. sector, volatility, dividend yield, etc.). The sector categories include Financials, Healthcare, Materials, Industrials, Consumer Discretionary, Consumer Staples and Utilities.


Akai Kaeru’s proprietary technology identifies groups of stocks where returns are unusually high/low. These groups are statistically robust and defined by simple criteria (e.g. Low volatility and High dividend). This information is then presented to the user through the interface shown in the figures below.

Akai Kaeru's XAI Dashboard: See it all in one place

image79

The top left plot contains the Group Bubble Chart. Each circle in this plot represents a collection of stocks which are similar in some way and where the returns are unusually high (green) or low (red). The opacity of the group tells us how important the x-axis attribute is in defining the groups. In this example, we can see a strong trend in which lower volatility is associated with higher returns. Additionally, the opacity of the groups tells us that volatility only becomes really important at its extremes.  


The user can select a group for further analysis. This is indicated by the circle with the black outline in the upper left of the bubble chart. The Group Summary in the top right panel show more detailed information. We can see that the selected group is described by low volatility (i.e. volatility <= 1.44) and high dividend (i.e. dividend >= 2.32).  

In the Probability Histogram panel (bottom center of the dashboard) we can see the distribution of returns for both the selected group (green) and full S&P 500 (gray). The bottom right panel shows the summary statistics for the selected group. Finally, the Feature Importance panel in the bottom left shows the importance of the different features both globally (gray) as well as the important features for the selected group (blue). Here we can see that although sector is the most important attribute globally, volatility is more important for the selected group. 

image80

This is the same dashboard, now showing the Group Detail in the top right panel (selected by pushing the respective tab on the top). The three cells for each feature are colored by the number of stocks in the low, mid, and high value interval. It indicates the attribute's most dominant (if any) value interval in the chosen group of stocks. We can see that the values for volatility are on the lower end while the values for dividend are on the higher end. 


The green bar in the last column shows for each feature the strength of its effect on return  which is positive for all features -- it would be colored red if the effect was negative.

Like to see our software in action?

Schedule a Demo

Copyright © 2020 Akai Kaeru, LLC - All Rights Reserved.

This website uses cookies.

We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.

Accept