前言
因为一些琐事，今天没有太多时间学习新的知识内容，故对已有资源进行了整理
LSTM 模型在量化交易中的应用汇总
Stock Clusters
Identifying stock clusters helps discover similar companies which can be useful for comparable analysis or pairs trading strategy. We can find similar clusters by estimating the inverse covariance (precision) matrix which can be used to construct a graph network of dependencies. The difference between opening and closing daily price was used to compute empirical covariance used to fit graph lasso algorithm to estimate sparse precision matrix. Affinity propagation was used to compute the stock clusters and a linear embedding was used to display high dimensional data in 2D.
链接 https://github.com/vsmolyakov/experiments_with_python
效果图如下
Plain Stock Close price Prediction via LSTM
This is a practice of using LSTM to do the one day ahead prediction of the stock close price. The dataset I used here is the New York Stock Exchange from Kaggle, which consists of following files:
- prices.csv: raw, as-is daily prices. Most of data spans from 2010 to the end 2016, for companies new on stock market date range is shorter. There have been approx. 140 stock splits in that time, this set doesn’t account for that.
- prices-split-adjusted.csv: same as prices, but there have been added adjustments for splits.
- securities.csv: general description of each company with division on sectors
- fundamentals.csv: metrics extracted from annual SEC 10K fillings (2012-2016), should be enough to derive most of popular fundamental indicators.
效果图如下
相关图书资料：《量化投资技术分析实战》