[AI] AI model for Taiwan Index Future(TXF/MXF)

Use DNN to find the best time to enter the market.

Technology

Python/Tensorflow

Motivation

In traditional quantitative trade, we use the index to classify whether to enter the market or not. And, classification is the most important application in ML/DL/AI. It worth a try to use AI way to find the best time to enter the market.

Description

Flowchart

Result

The profit trend is good enough, but it does not consider commission.

Train interval: 1999/3/12 ~ 2001/11/1

Test interval: 2001/11/1~2003/11/8

Profits of test interval: 5560 points (TXF: $200/point)

Trade # of test interval: 283

Profit:

Close of test interval