s3ntim3ntal

• 

  Logo of our project

Inspiration

My inspiration for this project is that my dad runs his very own software firm, and that means that clients usually have feedback for the services he provides. Since my dad doesn't have very much time while working his real job, I decided to develop a framework for something he could use in his real life to make customer feedback easier to analyze.

What it does

For now, the simple model I have trained can analyze the sentiment behind messages. It does this through NLP modules (natural language processing) to better understand the intent behind a review.

How we built it

I built this project by:

1. Finding a dataset to use (link in github page)
2. Referring to the dataset usage to train a machine learning model on the dataset
3. Ran it through the test data (91% accuracy)
4. Uploaded the model & a python file to interact with the model (also provided on github page)

Challenges we ran into

1. One of the challenges I ran into was issues training the model. Training the model is a complex process, and to someone who doesn't have too much working experience in the machine learning field, it really was a challenge, but I overcame it to make this wonderful piece of software.

Accomplishments that we're proud of

• Proud of the models high accuracy
• Proud of the speed and responsiveness of the model
• Proud of ease of usage
• Proud of the ease of deployment into a real-world scenario

What we learned

• How to train a model off a dataset
• How to write programs to query models and provide output
• Of course, building the website to link everything
• The uses of machine learning in an entrepreneurial perspective

What's next for s3ntim3ntal

• Train the model to work with shorter reviews, as it struggles with those
• Add more outputs, such as words that triggered certain sentiment, extracting the specific changes, and even an automatic response system using the same NLP modules.

Built With

• github
• google-colab
• kaggle
• python
• replit