GettingAIISentimental

What the Project Does:

Our program is an AI model which is designed to take a set of data and determine whether the sentences within the collection, demonstrate a positive or negative sentiment with great accuracy. Our program in particular is unique due to our "hybrid algorithm", in which we combined K Nearest Neighbors, and Support Vector Machine algorithms together. With each method being more accurate than the last, we were able to obtain precise results with great efficiency.

How We Built Our Model:

Using Python, our team built the model through Google Colab in which we were able to write and execute code through our browser and make modifications over a multitude of devices. We have used Sci-Kit Learn, MatPlotLib, Pandas, and various other libraries to amplify the accuracy of our program.

How Is Our Program Useful and Where Can It Be Applied:

Sentiment analysis is especially widespread in terms of its possibilities. It can largely be implemented in the customer service field, as well as predictive analytics. Our AI model can be used to gain insights into how customers feel about a product, and also help us identify and understand how business brand reputations develop.

How Users Can Get Started With the Project:

Our program can run on any platform that supports python and all that is required is to type in the authorization code given for the data set.

Challenges We Faced:

Working with large quantities of data proved to be a challenge, which we eventually overcame with dividing the procedure of data cleaning into smaller components.

Combining the two different methods of sentiment analysis to form our hybrid model tended to be very complex and convoluted especially as our team members did not already have previous experience with AI, but the effort and struggle eventually paid off when we developed our final finished product.

Another major challenge we had to overcome was dividing tasks among our team members and establishing when we would spend time on our project and when we would be resting. This was the first time any of our team members have done a hackathon and due to our inexperience there tended to be a level of disorganization, but despite that we were able to pull through, and now we’re all ready for hackathons in the future.