In this blog I am going to cover a prediction for sepsis using FastAPI framework and docker containerization. For this project a procees model called Cross-Industry Standard Process for Data Mining (CRISP-DM) is used.
Sepsis is a serious condition that occurs when the body’s immune system overreacts to an infection and causes widespread inflammation and organ damage.
It can lead to septic shock, which is a life-threatening drop in blood pressure that can result in organ failure and death.
Some of the risk factors for sepsis are:
- Being older than 65 or younger than 1
- Having a weakened immune system or a chronic medical condition
- Having a recent severe illness or hospitalization
- Having a device that goes into the body, such as a catheter or a breathing tube
- Having received antibiotics in the last 90 days
Developing a machine learning model for predicting sepsis positive or negative is good to identify sepsis at an early stage and provide timely and effective treatment.
Some studies have shown that machine learning models can outperform conventional methods of predicting sepsis mortality rates and have good accuracy and sensitivity. For further information about sepsis Click Here.
For This project I used some measurements those going to basement for prediction. These are : — Plasma glucose (PRG), Blood Pressure (mm Hg)Blood Work Result-1 (mu U/ml), Blood Work Result-2 (mm), Blood Work Result-3 (mu U/ml), Blood Work Result-4 (mu U/ml), Body mass index (weight in kg/(height in m), patients age (years), Valid insurance card
First work is to load necessary python packages like pandas, numpy, sklearn, matplotlib. This is the data’s used for prediction.
In below charts the first one explains Body Mass Index of persons. Half of the total is on over weight in BMI. As known optimal status is 18–24. The second graph tell us a distribution in age, which is clear that most of respondents are Young persons.
Now We reached to blood work results. As shown below first and second results are range to a specific boundaries. on the other hand the third one has unstable boundary.
Like blood result three the fourth also unstable. It is not helpfull get new insight. Insurance coverage is 70% which is good. This help peoples to get a continues medication without worrying about payment.