An Efficient Variant of Tf-idf Model to Enhance the Text-classification Performance
Year: 2022
Type: article
Abstract: Text classification and other text mining methods use as a baseline a classifier with term weight for the substantial increase in effectiveness. Different term weighting methods are used based on the words, documents and datasets. This study compares the effectiveness of various TF-IDF models and suggests a TF-IDF for text classification.. We proposed a mathematical definition of TF-IDF model based on the calculation and computation of usual TF-IDF, TF-IDF Variants, Term-frequency and Term-occurrence. We perform experiment on Amazon review dataset with different term weighting scheme, proposed approach perform better than on comparably to the other weighting scheme in terms of accuracy performance parameter using different classifier. It is shown in this paper that modifications to the original tf.idf weighting algorithms signif-icantly improve classification performance.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4307628331