Social Media Post Approval Prediction

Inspiration

Participated in TAMU Datathon for building collaboration, team building and technical skills within 24hrs.

What it does

Developed a model which predicts weather a AI generated post would be approved by the user.

How we built it

Developed model using scikit, pandas and NumPy modules. Where we used logistics regression, RandomForestClassifier and etc.

Challenges we ran into

biggest challenges was analyzing the dataset. As majority of data points were texts were a normal logistics regression model wouldn't be sufficient. Which meant using sentiment analysis to categorize those string into three labels: negative, neutral, positive which were based on the sentiment score of the text.

Choosing the appropriate model for the dataset was challenging as to get the best model, we compared accuracies of multiple models

Accomplishments that we're proud of

Developing a model from scratch which predicted with an accuracy of 0.6308 weather a AI generated post would be approved by user.

What we learned

scikit machine learning module, developing model using test and train datasets and analytical skills to develop the model.

What's next for Social Media Post Approval Prediction

One of them would be implementing tokens into the python script. developing more features for our model to train and predict with. Learn more about developing models and refactoring the code for future use.