Text data semantic space-time mode exploration method based on LDA model and LSTM network

Abstract

The invention discloses a text data semantic space-time mode exploration method based on an LDA model and an LSTM network. The text data semantic space-time mode exploration method comprises the following steps: (1) integrating a topic model, including theme generation, theme quality evaluation and theme dimension reduction projection; extracting semantics from the text data by using an LDA topicmodel, generating a topic model by iterating different parameters, performing quality evaluation on the topic model, and then selecting high-quality topics for integration so as to solve the influenceof the parameters on the model quality; (2) constructing a theme space-time body: converting time, space and text theme data in the text data into a cubic data structure; and (3) visual interaction and prediction: including a theme projection view, a time-space projection view and a mode comparison view; for providing visual interaction exploration for a theme space-time body, and being convenient for a user to explore a data result conveniently in a visual mode; and using an LSTM method to predict numerical change in a future time period.

Description

technical field [0001] This patent mainly relates to the fields of natural language processing and data visualization, and specifically relates to a method for structured representation of massive text data and topic model optimization. Background technique [0002] The amount of text data in the world has grown exponentially in recent years, which urgently requires people to mine new knowledge and new ideas from text data. From social media analysis to risk management and cybercrime protection, working with text data has never been more important. Since text data usually contains time and space information, spatio-temporal attributes are often added when text data is processed. [0003] Among the works dealing with the distribution of text data in space and time, there are many works that focus on finding keywords in text data. A common approach is to analyze text data to detect associated events that occur at specific times and locations, identifying events from text gro...

AI Technical Summary

Problems solved by technology

[0004] The above-mentioned works mainly use the text data of social media, but there are some shortcomings and deficiencies in these works: first, the above-mentioned work is to set the parameters of the topic model in advance, and use the trained topic model to extract semantics from the data of the present invention, because the topic model Sensitive to parameters, this cannot guarantee the quality of the topic model, thus affecting the ability of the model to extract semantics

Secondly, the above work mainly uses static data, and there is no requirement for data processing speed and query speed. When faced with massive text data, it is often impossible to respond in a timely manner.

Finally, the above work only provides a query function when displaying data, and cannot add the user's decision to the result

Images