Method for automatically identifying document research on the basis of text

Abstract

The invention discloses a method for automatically identifying document research on the basis of a text. The method is characterized in that: firstly, on the basis of a CRF (Conditional Random Field) model, modeling is carried out on a small quantity of labeled data; then, unlabeled data is predicted, and parts of data are selected from a majority of prediction labeling set as little as possible to carry out artificial labeling; then, a labeled result is added into original corpora to carry out modeling again, the process is subjected to proper iteration to obtain a final model, and the model can be used for extracting the research objects of scientific and technical literatures. The method comprises the following steps that: S1: obtaining a scientific and technical literature title, and carrying out primary labeling; S2: carrying out standardization processing on data; S3: carrying out model characterization extraction; S4: training the data; S5: extracting parts of unlabeled data, and carrying out labeling; and S6: estimating model accuracy. Artificial assessment is optimally imported into a machine learning model, the improvement effect of the machine learning model can be effectively improved and artificial labeling cost can be saved as far as possible.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a text-based method for automatically identifying literature research. Background technique [0002] A scientific paper is written by the author to summarize and refine the presentation of the research work. A scientific paper generally includes different research elements, such as research background, research object, research process, research method, research conclusion, etc. The research object refers to the main content of the paper. The core subject of the research objective can efficiently and clearly locate the focus of the corresponding article, including attribute examples such as objective things, theories, events, processes, and relationships. The extraction of research objects can display the main research objectives of the paper in an intuitive form. It helps researchers to quickly grasp the relevant information of this object, and to retrieve and co...

AI Technical Summary

Problems solved by technology

[0002] A scientific paper is written by the author to summarize and refine the presentation of the research work. A scientific paper generally includes different research elements, such as research background, research object, research process, research method, research conclusion, etc. The research object refers to the main content of the paper. The core subject of the research objective can efficiently and clearly locate the focus of the corresponding article, including attribute examples such as objective things, theories, events, processes, and relationships. The extraction of research objects can display the main research objectives of the paper in an intuitive form. It helps researchers to quickly grasp the relevant information of this object, and to retrieve and compare relevant research content conveniently. There are many existing processing methods, including rule-based methods. This method has achieved certain results, but due to natural language Due to the diversity of sentence patterns, this method cannot cover all the rules in the extraction of research objects, and there are too many omissions, and it cannot be updated in real time, so the flexibility is poor. However, the extraction process using statistical learning methods often introduces uncertain factors and imports recall. High and low accuracy, therefore, the practicability of pure rules or statistics is very limited

Images