pooyaphoenix / hystoclass

Topic classification of social network content: combining text and graphical features using transformer-based architecture

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

hystoclass

Topic classification of social network content: combining text and graphical features using transformer-based architecture

Dataset: https://huggingface.co/datasets/pooyaphoenix/hystoclass

Result:

table

The results show that the model based on ensemble classification has more performance. The final result is that the graphical features in the data such as background color, text color, and font can increase the performance of text classification.

About

Topic classification of social network content: combining text and graphical features using transformer-based architecture


Languages

Language:Jupyter Notebook 100.0%