An App-Based Recommendation System for Cognitive Exercises and Physical Training Programs

The Omdena team delivered a prototype of a Recommendation System for the NESTRE app, incorporating rule-based and machine learning models for content recommendations, utilizing synthetic data during the prototype stage, and future plans to optimize the model with real-world data.

The problem

NESTRE is the first NEuro-STREngth company to leverage the science of neuroplasticity to successfully apply a proprietary and pioneering training method in the areas of brain health, mental wellness, and cognitive performance. NESTRE’s first digital solution is a neuro-strength training platform that provides cognitive assessments and training to users. NESTRE’s mission is to help people get better by providing access to affordable solutions and instilling hope for betterment through proven science.

The project outcomes

After an 8-week development period, the Omdena team successfully delivered a prototype of a Recommendation System for the Daily Workout and Navigational Map content of the NESTRE app. The system utilizes both rule-based and machine-learning approaches for making recommendations. Specifically, the rule-based model is used for content recommendations related to the Daily Workout, while the Navigational Map content recommendations employ machine learning models.

During the prototype stage, the model was trained using synthetic data generated by the Omdena team. This data is not intended for production use and was solely utilized for training and testing purposes, allowing us to observe the system’s behavior in various scenarios, runtime performance, scalability of machine learning models, and optimization techniques.

To facilitate interaction with the models and display the results, a simple web front-end was developed to work in conjunction with our Django API.