Created on 7th August 2022
•
We live in the age of digital life where everything is done online. We interact with people on the Internet more often than in real life. But most of the time these interactions don't go well. We've seen people harassing each other, passing negative comments, and talking about violence too often. This has led to a very chaotic digital experience. Through our project, we are trying to minimize the toxic ingredients of the digital world. Our web app simply tries to generate as friendly replies as possible to a given sentence. It checks if the given sentence is harassing and returns something that's more friendly and peaceful. We have worked with Natural Language Processing to identify the sentiments of the sentence and get the desired result
The major challenge we had was our inexperience with AI and Natural Language Processing. We were a team of only two people with some knowledge of python and web applications. We thought we had a great idea to implement but limited knowledge to apply it.
The first problem we ran into was preprocessing the data we had. We got our dataset from Kaggle and it was really messed up. We had absolutely no clue how could we clean that data but after a lot of google searches we somehow managed to find a workaround for it. Another major challenge we ran into was connecting our Machine learning model to the web which we still couldn't figure out. Since we had limited knowledge in this area we couldn't make it work in the given time interval.
Another challenge was the time zone for both of us. We had to pick a suitable time for both of