Skip links

Predict health alert based on patient’s vital parameters

Client

Pharmaceutical company from Germany

Scroll down

Technologies used

Postgre SQL

Python

Challenges

Understanding the data sources and finalizing true source for each parameter

Identify best definition to predict positive patients

Imbalanced dataset as positive patients are much lesser than negative patients in overall cohort

Solutions

Data Preparation

Handling missing values

Handling imbalanced dataset

Feature Engineering

Model Training

Model Evaluation and testing

Identifying best model

Benefits

Understanding vital parameters for positive and negative patients

Predicting disease before onset time

Saves patient from emergency situation