-
-
Insights Expo
Inspiration
Insights expo wanted to pick a solution that would be relevant to our respective countries, Nigeria and India. Salisu, being a med student, recognized that doctors faced a lot of challenges in providing treatment to patients in low-income areas due to dietary deficiencies in their patients’ diets. After some deliberation, we decided malnutrition is a very critical point of interest and should be given more importance and awareness than it currently has.
What it does
Our model uses relevant parameters to predict malnutrition rates in an area and the effect of malnutrition likely to be more prevalent between being overweight, stunted, or wasting. Using these insights, relevant authorities can tailor malnutrition alleviation strategies specific to their communities.
How we built it
We built the model by using two different ML models and comparing which the better ones are. We used Linear regression and Random Forests and found Random Forest was the better algorithm. This was classified based on different types of loss methods like MSE and RMSE. We used a test case to predict Stunt, Overweight, and Wasting. This worked well as a prediction and with more adequate data, this can work perfectly well to predict.
Challenges we ran into
The dataset we used was originally column-wise and we had to do some machinations to make it row-wise. Also, it contained thousands of null values.
Accomplishments that we're proud of
We’re proud of the fact that we didn’t let any challenges we came across discourage us. We offered each other help and most importantly, we learned together and conquered all difficulties together. The model was also built for an important yet difficult problem our society faces and hence taking it on our shoulders was in itself an accomplishment.
What we learned
We learned a lot about IBM Z and we were able to deepen our knowledge on data science concepts we didn’t know well before this hackathon such as visualization. We learned about ensemble models, model evaluation metrics and so much more. This project was a continuous learning experience.
What's next for Team-49: Malnutrition Prediction In Children Under 5
We would like to incorporate recommendations into our project for communities to consult when devising a strategy against malnutrition.
Log in or sign up for Devpost to join the conversation.