Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

News sub-event prediction method and apparatus based on depth learning technology

A prediction method and sub-event technology, applied in the field of intelligence analysis, can solve the problems of not considering the sequence structure, not considering the use of topic information, and only predicting.

Inactive Publication Date: 2017-09-01
TSINGHUA UNIV
View PDF1 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] However, the above-mentioned news sub-event prediction techniques all have the following problems: 1. Artificial features are required to represent events and can only predict events that exist in the training set; 2. The sequence structure of events at different granularities is not considered: Words describing a specific sub-event form a sequence, and sub-events belonging to the same event form a sequence; 3. No consideration is given to utilizing topic information existing in existing sub-events

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
  • News sub-event prediction method and apparatus based on depth learning technology
  • News sub-event prediction method and apparatus based on depth learning technology
  • News sub-event prediction method and apparatus based on depth learning technology

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0020] The specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0021] The present invention proposes a new model for sub-event prediction, that is, Hierarchical Long Short-Term Memory (HLSTM) hierarchical long-short-term memory model. The input of the model is text describing observed sub-events (such as news headlines), and the output is a short text describing future sub-events. HLSTM is a hierarchical LSTM structure with two layers. The first layer is used to model sub-events. It sequentially takes each word in the description text of the sub-event as input, and finally represents the sub-event as a vector. The second layer is used to model the observed sequence of sub-events, while considering the topic information to enhance t...

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 present invention provides a news sub-event prediction method and apparatus based on the depth learning technology. The method comprises: step 1: carrying out representation learning on the sub-event by using the LSTM depth learning technology based on the pre-processed large-scale sub-event sequence data; Step 2, carrying out representation learning on the sub-event sequence based on the sub-event representation learning and sub-event subjects; and step 3, predicting the next sub-event by using the sub-event sequence representation obtained in the step 2. According to the method provided by the present invention, the sub-event prediction function can be effectively realized, the sub-event not existing in the training corpus can be predicted, and the better prediction effect can be obtained.

Description

technical field [0001] The present invention relates to the technical field of intelligence analysis, and more specifically, to a method and device for predicting news sub-events based on deep learning technology. Background technique [0002] At present, a news event refers to one or a series of events that occur at a specific time and place, and a news event is usually composed of a series of ordered sub-events. For example, an earthquake event usually includes a series of sub-events such as earthquake occurrence, earthquake casualties, earthquake financial losses, post-earthquake rescue, post-earthquake donation assistance, and post-earthquake disaster area reconstruction. A sub-event is usually described by a news report. [0003] So far, thousands of news events and their sub-events have been reported and recorded. It is possible to predict the next sub-event of a series of sub-events by exploiting large-scale historical events. Through the above-mentioned scientific...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06F17/27G06N3/08
CPCG06F40/30G06N3/08
Inventor 李涓子胡琳梅
Owner TSINGHUA UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products