Convolutional neutral network-based attribute extraction method

Abstract

The invention discloses a convolutional neutral network-based attribute extraction method. The method comprises the following steps of (1) constructing an external knowledge library; (2) obtaining text data; (3) obtaining attribute-containing sentences by using a remote supervision method; (4) obtaining the sentences by utilizing a word vector method and performing vectorization; and (5) inputting the sentences to a convolutional neutral network, and performing training and classification. According to the method, the attribute-containing candidate sentences are extracted from a non-structured text data set based on artificially defined mapping by utilizing the external knowledge library in combination with remote supervision and convolutional neutral network models, and sentence classifications are classified in combination with the convolutional neutral network model, so that an attribute extraction task is finished.

Description

technical field [0001] The invention relates to text feature extraction and attribute extraction, in particular to an attribute extraction method based on a convolutional neural network. Background technique [0002] The world today is in an era of information explosion. The popularity and rapid development of the Internet have produced massive information resources. These resources are of great significance to the development of science and technology. The scientific community needs to extract the basic materials of scientific research from them, and the industry needs to dig out potential business opportunities from them. Therefore, how to use these Internet information resources has become the mainstream direction of scientific and technological research in recent years. one. [0003] Although the amount of information resources on the Internet is huge, these resources often lack structured features. Structured data refers to row data, data that can be expressed in a tw...

AI Technical Summary

Problems solved by technology

[0005] The traditional attribute extraction technology has many disadvantages. First of all, whether it is based on rules or classification algorithms, it requires more manual intervention, such as rule-based rule design and classification-based data labeling and feature design. The cost of manual intervention is Expensive, and requires the labeling of professionals to obtain more authoritative artificial data. At the same time, manual work will also bring certain errors, and errors will continue to accumulate in subsequent algorithms, which will eventually lead to excessive deviation of the results; secondly, The training data set of this type of algorithm is limited in a certain field, that is, it is not universal. For example, the attribute extraction classifier trained on sports news cannot be well used in other news; usually the above algorithm obtained The effect is generally not ideal, because the artificially designed rules in the rule-based method are limited, and the label data of the classification-based method is also limited, and this method is more dependent on the quality of the artificially designed features.

Images