Solution to fight Chicago Gang Violence with AI
My story at Omdena Voice4Impact to find a solution for a very serious problem in Chicago. My team contributions helps to take countermeasures on negatively impacted tweets which may prevail Gang Violence.
Gun violence in Chicago threats everything and it is rapidly increasing .We at Omdena partnered with Voice4Impact to predict potential threats from online chatter, posts using AI
Some examples for violent behavior we found in our data set
“ Shot a nigga took a shower wit the bleach”
“🤣🤣😭 Can’t Get Over The Fact That Nigga Said He Was Gon Shoot Some Shit Up.”
I was part of sentimental analysis team to predict whether the tweets are threatening. Most of the tweets are non threatening , it is a imbalanced dataset. We tried different techniques like undersampling, transfer learning from existing word embedding algorithms, vader sentiment analysis. Tried models like SVC, XGBoost, LSTM, Logistic Regression,ensemble models.We split the total data set into sub train and test sets to over come imbalance problem.
Final result of LSTM + Word2Vec model
Collaborative approach at Omdena is like “ Little drops of water Make the Mighty Ocean , I am delighted to be a little drop ”
Omdena helps by connecting us to solve some of the greatest challenges on the planet. I love to be a part of greater good.
If you want to be part of the #AIforGood movement, join our global community as an (aspiring) data scientist or as an organization.
