Write It Right

Screenshot of website

Inspiration

We've learned from our experiences at school that essay grading today is highly inefficient and subjective. With machine learning taking on a rapidly increasing number of roles in our daily lives, we decided to implement an artificial intelligence system through machine learning to automate the essay grading process.

What it does

Our machine learning model is built into a simple website. Students or teachers can paste in an essay into a text box, and it is run through a robust model trained on thousands of past student essays to determine a score. Currently, it is evaluated as pass or fail to allow students to determine whether they need to spend more time on their essay. We will soon expand to produce a numerical grade on a scale from 1 to 12.

How we built it

The core of our model is built in Python. We used the library NLTK for the natural language processing functionality to analyze the text. We trained a decision tree model on the Kaggle Hewlett Foundation essay grading dataset for classification. All Python was done in a Jupyter Notebook through Anaconda. Our website is built in HTML, CSS, and JavaScript. We will integrate the machine learning model into an API to connect it to the website.

Challenges we ran into

We ran into some difficulties in training our model initially, as our accuracy was far lower than we expected. Through adding additional features for analysis and experimenting with a variety of model structures, we were able to greatly improve the accuracy.

Accomplishments that we're proud of

We are proud to have figured out how to successfully train a robust model despite only one team member having prior experience in machine learning.

What we learned

We greatly expanded our knowledge of machine learning, natural language processing, and web development.

What's next for Write It Right

Our number one goal is to provide more precise feedback in the form of a grade from 1 to 12. We also want to improve our overall accuracy, and use an API to integrate the model into our website.