Scribd
Upload
27 views

Report

Team ARIMA presented a machine learning model to predict health risks for pregnant patients. They built a random forest classifier using data on patient attributes like age, blood pressure, and glucose levels. The model achieved 90% accuracy in predicting low, medium, and high risk levels. They deployed the model through a web app using Flask and Heroku to allow clinicians to input patient data and receive risk predictions.

Uploaded by

Anchal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
27 views

Report

Team ARIMA presented a machine learning model to predict health risks for pregnant patients. They built a random forest classifier using data on patient attributes like age, blood pressure, and glucose levels. The model achieved 90% accuracy in predicting low, medium, and high risk levels. They deployed the model through a web app using Flask and Heroku to allow clinicians to input patient data and receive risk predictions.

Uploaded by

Anchal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 7

HDSC Winter’23 Premiere Project Presentation:

Predicting Health Risks For Pregnant Patients


A Project by Team ARIMA

pregnant woman consulting in gynaecology office

A high-risk pregnancy is one that threatens the health or life of the mother or her
foetus. It often requires specialized care from specially trained providers.

Some pregnancies become high risk as they progress, while some women are at
increased risk for complications even before they get pregnant for a variety of
reasons.

Early and regular prenatal care helps many women have healthy pregnancies and
deliveries without complications.

High-risk pregnancy is a term that can denote a wide variety of common conditions.
Many of them are related to pre-existing conditions you may have had before
becoming pregnant or conditions you may have developed while pregnant or during
delivery.
Aims and Objectives
The aim of this project is to build a machine learning model that predicts the risk
level in pregnant patients. This dataset has different factors influencing patient
health, we will classify the patients with a low, medium or high risk of having
pregnancy complications.This can be achieved by training a machine learning model
with past records of some patients having various features.

Methodology
The whole process of a machine learning project is explained in this flowchart
starting with data collection along with its cleaning and moving with model training
and evaluation till model deployment.

flowchart showing the whole process of ml project


Data Description
An open-source data obtained from Kaggle was used, it can be accessed via the link
below:

mailto:https://www.kaggle.com/datasets/csafrit2/maternal-health-risk-data?
resource=download

Data has been collected from different hospitals, community clinics, maternal health
cares through the IoT based risk monitoring system.

 Age: Age in years when a woman is pregnant.


 SystolicBP: Upper value of Blood Pressure in mmHg, another significant
attribute during pregnancy.
 DiastolicBP: Lower value of Blood Pressure in mmHg, another significant
attribute during pregnancy.
 BS: Blood glucose levels is in terms of a molar concentration, mmol/L.
 HeartRate: A normal resting heart rate in beats per minute.
 Risk Level: Predicted Risk Intensity Level during pregnancy considering the
previous attribute.

Data Pre-processing
 The dataset has 1014 rows and 7 columns of dataentries.
 There are no missing values in the data.
 Number of rows with duplicates are 562.
 Most of the columns are numeric.

 Converted column RiskLevel with string data type to numeric data type to
make them fit for analysis and modelling.
 To get familiar to the dataset, we used describe() function and gathered some
information.

Exploratory Data Analysis


We did some data analysis including Univariate and Bivariate analysis to understand
our data better and here it is shown below:

 This distribution of risklevel is shown here in pie chart which is equally


distributed among all risk levels.

 In these plots, we will see the dependency of age against the column
risklevel.
 This plot shows that with increase in age, risk level increases.
 This plot shows the distribution of heart rate with risk level.

 From the below heat map, we see how each feature are related with each
other.
Model Training And Evaluation
As our problem statement is classification based, so we tried different classification
algorithms in order to find the model with highest accuracy.
About eight (8) machine learning algorithms were evaluated on the resulting dataset.
The algorithms include:
KNN, DecisionTreesClassifier, AdaBoostingClassifier, RandomForest,
BaggingClassifier, XGBClassifier, CalibratedClassifierCV and GaussianNB.

According to our classification models in this study, the best-performing model was
the RandomForestClassifier so we stick to it.

In order to increase its accuracy , we did hyperparameter tuning using


GridSearchCV as well and got the accuracy of parameterised model of around 90%.

Model Deployment
The deployment was done using the flask framework in python programming
language and was hosted using the Heroku platform. The saved model and
transformer were deployed alongside in order to transform new and subsequent
“facts” and “issue_area” returned in the form from the frontend of the project to make
predictions for the input given by the user of the web app. Here are the links;
Heroku:
Link to repo:
Conclusion
Recommendations
References
 Using Machine Learning to Predict Complications in Pregnancy: A Systematic
Review,
mailto:https://www.frontiersin.org/articles/10.3389/fbioe.2021.780389/full
 Pregnancy Outcome Prediction study,
mailto:https://en.wikipedia.org/wiki/Pregnancy_Outcome_Prediction_study

 MODEL FOR PREDICTING RISK LEVELS IN MATERNAL HEALTHCARE,


mailto:https://ijariie.com/AdminUploadPdf/
MODEL_FOR_PREDICTING_RISK_LEVELS_IN_MATERNAL_HEALTHCAR
E_ijariie18831.pdf

Team Members

You might also like