Stock price data prediction method and device

A data prediction and stock price technology, applied in the computer field, can solve the problems of ignoring long-memory characteristics, affecting the trend of stock prices, and low accuracy of stock price prediction results, so as to solve long-term dependencies and improve accuracy.

Inactive Publication Date: 2022-02-15
深圳希施玛数据科技有限公司
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] Stock prices fluctuate in real time, with great instability and randomness. In the process of stock trading, stock selection and purchase behaviors are often based on people's subjective decisions or emotionally based on the rise and fall of stock prices. Therefore, Investor public opinion will affect the follow-up price trend of the stock
[0003] At present, when predicting future stock prices, the impact of current investor public opinion on current stocks is generally used, ignoring the long-memory characteristics between stock price time series and investor public opinion time series, making the accuracy of stock price prediction results low.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Stock price data prediction method and device
  • Stock price data prediction method and device
  • Stock price data prediction method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] see image 3 , image 3 It is a structural schematic diagram of a stock price prediction model provided in the embodiment of this application. Such as image 3 As shown, the stock price prediction model includes a sentiment generation model and a prediction model.

[0047] Among them, the input of the sentiment generation model is input data (investor's comment data), and the output is text sentiment and investor sentiment. The emotion generation model can include different Bi-LSTM models based on the Attention mechanism, such as image 3 As shown, the sentiment generation model includes a first Bi-LSTM model, a second Bi-LSTM model and a Softmax layer, the first Bi-LSTM model is used to obtain text sentiment features, and the second Bi-LSTM model is used to obtain investment The Softmax layer is used to concatenate and calculate text sentiment features and investor sentiment features to obtain the final text sentiment and investor sentiment.

[0048] The training ...

Embodiment 2

[0054] see Figure 4 , Figure 4 It is a schematic flowchart of a stock price data prediction method provided in the embodiment of this application. Such as Figure 4 As shown, the method includes the following steps:

[0055] S410. Obtain a target data set and a target stock price, the target data set includes multiple comment data published by multiple first investors on the target stock within the i-th period, and the target stock price is The average stock price of the target stock within the period, and the i is a positive integer greater than 1.

[0056] Wherein, the above-mentioned target data set may be the comment data published by investors on the target stock automatically obtained from the stock investment application program or page, and the comment data may include at least one of text, emoticon, and image. The cycle time can be set independently by the user, such as 3 hours, one day, two days, one week, two weeks, one month, etc.; it can also be set by the s...

Embodiment 3

[0070] An embodiment of the present application provides a method for training a stock price prediction model, and the stock price prediction model may be the stock price prediction model in Embodiment 2. This training method can be used for image 3 The stock price prediction model is trained, and the stock price prediction model includes an emotion generation model and a prediction model. Since the emotion generation model and the prediction model are both Bi-LSTM models, and the output of the emotion generation model is the input of the prediction model, it can be used The training data trains both the sentiment generation model and the predictive model.

[0071] see Figure 5 , Figure 5 It is a training method for an emotion generation model provided in the embodiment of this application, and the emotion generation can be the emotion generation module in the first embodiment. This training method can be used as image 3 The structural implementation of the stock price...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The embodiment of the invention provides a stock price data prediction method and device. The method comprises the following steps: obtaining a target data set and a target stock price, wherein the target data set comprises a plurality of pieces of comment data published by a plurality of first investors for a target stock within the ith cycle time, and the target stock price is the average stock price of the target stock in the ith cycle time; and inputting the target data set and the target share price into the share price prediction model to obtain an output result, the output result including a predicted share price, the predicted share price being an average share price within the predicted (i + m) th cycle time, and m being a time sliding window. According to the invention, the stock price is predicted through the endogenous recursive stock prediction model, the problem of long-time dependence of stock price prediction can be solved, and the accuracy of the stock price prediction result is improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a stock price data prediction method and device. Background technique [0002] Stock prices fluctuate in real time, with great instability and randomness. In the process of stock trading, stock selection and purchase behaviors are often based on people's subjective decisions or emotionally based on the rise and fall of stock prices. Therefore, Investor public opinion will affect the subsequent price trend of the stock. [0003] At present, when predicting future stock prices, the impact of current investor public opinion on current stocks is generally used, ignoring the long memory characteristics between stock price time series and investor public opinion time series, making the accuracy of stock price prediction results low. Contents of the invention [0004] The embodiment of the present application provides a stock price data forecasting method and device, whic...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06Q10/04G06Q40/04G06N3/04G06N3/08
CPCG06Q10/04G06Q40/04G06N3/08G06N3/044
Inventor 穆旖旎张中岩
Owner 深圳希施玛数据科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products