Subject Domain Extraction and Classification Model

Time

From 2017-08 to 2017-12

Project Introduction

Aiming at the problem that the growing Chinese scientific and technological literature cannot be classified automatically. Manual reading and tagging will increase labor costs, and manual methods rely too much on the professional skills of practitioners.

Procedure

Therefore, we automatically classify the scientific and technological literature by designing the deep learning network, and the reference class is the CNKI. In other words, the project is an application of “Text Multi-label”. By completing the text multi-label network, we test and set different thresholds so that the distribution of scientific and technical documents is as close as possible to the original distribution, so as to improve the whole system.