scholarly journals Construction of Quantitative Transaction Strategy Based on LASSO and Neural Network

2017 ◽  
Vol 4 (4) ◽  
pp. 134 ◽  
Author(s):  
Xu Wang ◽  
Zi-Yu Li ◽  
Jia-Yu Zhong
Keyword(s):  
Neural Network ◽  
Stock Price ◽  
Trading Strategy ◽  
Stock Index ◽  
Good Effect ◽  
Stock Index Futures ◽  
Expected Return ◽  
Index Futures ◽  
Statistical Field ◽  
Input Variables

Since the establishment of the securities market, there has been a continuous search for the prediction of stock price trend. Based on the forecasting characteristics of stock index futures, this paper combines the variable selection in the statistical field and the machine learning to construct an effective quantitative trading strategy. Firstly, the LASSO algorithm is used to filter a large number of technical indexes to obtain reasonable and effective technical indicators. Then, the indicators are used as input variables, and the average expected return rate is predicted by neural network. Finally, based on the forecasting results, a reasonable quantitative trading strategy is constructed. We take the CSI 300 stock index futures trading data for empirical research. The results show that the variables selected by LASSO are effective and the introduction of LASSO can improve the generalization ability of neural network. At the same time, the quantitative trading strategy based on LASSO algorithm and neural network can achieve good effect and robustness at different times.

Trading Strategy for Stock Index Futures Based on Disposition Effect

2014 ◽  
Vol 22 (1) ◽  
pp. 25-44
Author(s):  
Seung Hyun Oh ◽  
Sang Buhm Hahn
Keyword(s):  
Investment Strategy ◽  
Trading Strategy ◽  
Capital Gains ◽  
Stock Index ◽  
Expected Returns ◽  
Stock Index Futures ◽  
Investment Strategies ◽  
Positive Relation ◽  
Index Futures ◽  
The U.S

Grinblatt and Han (2005) argued that unrealized capital gains and expected returns are positively related in the U.S. stock markets. This study investigates the possibility of developing investment strategies for stock index futures using the positive relation. Probing the trading data of futures on KOSPI200 during the period of 1995~2013, several interesting results are obtained. First, the strategy of building long positions when the unrealized capital gain is greater than zero produces positive profit which is statistically significant. Second, the profitability of this strategy during December is significantly positive while the profitability during January is insignificant. Third, the strategy generates positive profit during the second half year while the profitability of the first half year is insignificant. These results imply that it is possible to develop investment strategy by extracting some information from the unrealized capital gains.

scholarly journals Futures Trend Strategy Model Based on Recurrent Neural Network

2018 ◽  
Vol 5 (4) ◽  
pp. 95
Author(s):  
Ru Zhang ◽  
Chenyu Huang ◽  
Shaozhen Chen
Keyword(s):  
Neural Network ◽  
Recurrent Neural Network ◽  
Futures Market ◽  
Stock Index ◽  
Stock Index Futures ◽  
Investment Performance ◽  
Index Futures ◽  
Model Based ◽  
Strategy Model ◽  
Global Futures

In recent years, quantitative investment has been widely used in the global futures market, and its steady investment performance has also been recognized by domestic futures investors. This paper takes the CSI-300 stock index futures as the research object and constructs a futures trend strategy model based on recurrent neural network. Furthermore, this paper back tests the strategy at different periods, different transaction costs and different parameters. The results show that the strategy model has strong profitability and robustness.

scholarly journals Bollinger Bands Trading Strategy Based on Wavelet Analysis

2018 ◽  
Vol 5 (3) ◽  
pp. 49
Author(s):  
Shaozhen Chen ◽  
Bangqian Zhang ◽  
GengJian Zhou ◽  
Qiaoxu Qin
Keyword(s):  
Wavelet Analysis ◽  
Noise Reduction ◽  
Risk Ratio ◽  
Trading Strategy ◽  
Trading Strategies ◽  
Stock Index ◽  
Stock Index Futures ◽  
Income Risk ◽  
Index Futures ◽  
Quantitative Trading Strategies

With the popularization of the concept of quantitative investment and the introduction of stock index futures in China, the research on the quantitative trading strategies of stock index futures is emerging gradually. This paper takes the CSI 300 stock index futures as the research object and sets up the Bollinger Bands trading strategy to test it, while considering the factors such as returns, retracement and income risk ratio, etc. Furthermore, the paper uses the wavelet noise reduction to process the data of price and the Bollinger Bands trading strategy to test the processed data. Compared with the results of the first test, the Bollinger Band trading strategy based on wavelet analysis has greater returns, less risk and better applicability.

scholarly journals Research on Quantitative Trading Strategy Based on Neural Network Algorithm and Fisher Linear Discriminant

2017 ◽  
Vol 9 (2) ◽  
pp. 133 ◽  
Author(s):  
Zi-Yu Li ◽  
Yuan-Biao Zhang ◽  
Jia-Yu Zhong ◽  
Xiao-Xu Yan ◽  
Xin-Guang Lv
Keyword(s):  
Neural Network ◽  
Bp Neural Network ◽  
Trading Strategy ◽  
Stock Index ◽  
Stock Index Futures ◽  
Fisher Linear Discriminant ◽  
Linear Discriminant ◽  
Price Forecast ◽  
Network Algorithm ◽  
Neural Network Algorithm

Based on the trend background of financial development in China in recent years, and statistical analysis of trend line, this paper establishes the quantitative trading strategy through the BP Neural Network Algorithm and the Fisher Linear Discriminant. Firstly, the data is linearly regressed into equal-length trend lines and the slope is fuzzified to build the matrix of upward trend and downward trend. And then use BP Neural Network Algorithm and Fisher Linear Discriminant to carry on the price forecast respectively and take transaction behavior, and correspondingly we take Shanghai and Shenzhen 300 stock index futures as an example to carry on the back test. The result shows that, firstly, the initial price trend is well retained by fitting; secondly, the profitability and risk control ability of the trading system are improved through the training optimization of Neural Network and Fisher Linear Discriminant.

Sign in / Sign up

Export Citation Format

Share Document