# Mimicking Human Process: Text Representation via Latent Semantic Clustering for Classification

## TL;DR

This paper introduces a novel text representation method that clusters words based on their latent semantics to improve classification performance, validated through experiments on multiple benchmarks.

## Contribution

It proposes a new semantic clustering-based text representation scheme that enhances classification accuracy over traditional methods.

## Key findings

- Effective on five classification benchmarks
- Visualization confirms meaningful semantic clustering
- Improves semantic expression for text classification

## Abstract

Considering that words with different characteristic in the text have different importance for classification, grouping them together separately can strengthen the semantic expression of each part. Thus we propose a new text representation scheme by clustering words according to their latent semantics and composing them together to get a set of cluster vectors, which are then concatenated as the final text representation. Evaluation on five classification benchmarks proves the effectiveness of our method. We further conduct visualization analysis showing statistical clustering results and verifying the validity of our motivation.