Skip to content

Latest commit

 

History

History
15 lines (8 loc) · 841 Bytes

File metadata and controls

15 lines (8 loc) · 841 Bytes

Automatic Evaluation of Textual Descriptive Answers

This repository contains the implementation of Automatic Evaluation of Textual Descriptive Answers on the famous Automated Essay Scoring (AES) dataset downloaded from here on Kaggle.

The Sequential API from Keras has been used as a classifier model. The following flow diagram shows the model.

Keras Plot

The model performs quite well in terms of Cohen's Kappa score and achieves an average score of about 0.96

A line graph showing 30 randomly selected questions and their predicted counterparts by the model is shown below:

Line Plot

NOTE: The data used is in un-normalized form i.e., the target scores are not in a constant range.