Research Article
Text Classification Feature Extraction Method Based on Deep Learning for Unbalanced Data Sets
@INPROCEEDINGS{10.1007/978-3-030-67871-5_29, author={Li Lin and Shu-xin Guo}, title={Text Classification Feature Extraction Method Based on Deep Learning for Unbalanced Data Sets}, proceedings={Advanced Hybrid Information Processing. 4th EAI International Conference, ADHIP 2020, Binzhou, China, September 26-27, 2020, Proceedings, Part I}, proceedings_a={ADHIP}, year={2021}, month={2}, keywords={Deep learning Unbalanced data sets Text features Classification and extraction}, doi={10.1007/978-3-030-67871-5_29} }- Li Lin
Shu-xin Guo
Year: 2021
Text Classification Feature Extraction Method Based on Deep Learning for Unbalanced Data Sets
ADHIP
Springer
DOI: 10.1007/978-3-030-67871-5_29
Abstract
In order to fully realize the classified search of text data information, a text classification feature extraction method for imbalanced data sets based on deep learning is proposed. With the help of trestle automatic encoder and depth confidence network, the preliminary definition of text semantic category conditions is completed, and the text semantic classification processing based on depth learning algorithm is realized. On this basis, pre-processing and debugging of text parameters are implemented, and the dimensionality reduction standards related to the text features of the data set to be extracted are established through the expression of the characteristic behavior. The experimental results show that with the application of the new classification feature extraction method, the number of correctly classified documents starts to increase substantially, which meets the practical application requirements for the classification and search of text data information.
