- Problem Statement
- Solution
- Features
- Machine Learning Models
- Installation
- Usage
- Contributing
- License
All newborns cry and get fussy sometimes. It is normal for a baby to cry for 2-3 hours day for the first 6 (six) weeks. During the first 3 (three) months of life, they cry more than at any other time (Pitone, 2023). The impact of crying on paretns can trigger feelings of anxiety, depression, helplessness, anger and frustration, which negatively affect bonding with the baby.
Our solution is a smart crib monitoring system that ensures the safety and comfort of babies. It monitors the baby's surroundings including humidity, and temperature. Additionally, it classifies the baby's cry using machine learning algorithms.
- Environment Monitoring: Monitors the baby's surroundings for optimal comfort. It checks for parameters like humidity and temperature.
- Cry Classification: Classifies the baby's cry using machine learning algoritms. It classify into 5 categories: Belly Pain, Burping, Discomfort, Hungry, and Tired.
We are using Support Vector Machines (SVM) and Random Forest algorithms for the cry classification feature. These models were chosen for their effectiveness in pattern recognition and adaptability to different data distributions.
Our project has benefited from open-source codes and libraries. We would like to acknowledge the following:
- Github Repository:
Machine Learning Systems - KWS Feature Engineering. (n.d.). https://harvard-edge.github.io/cs249r_book/contents/labs/shared/kws_feature_eng/kws_feature_eng.html