Neonatal-Weight-Prediction-Model-with-Inferential-Statistics-in-R

🎯 Project Objective:

This project aims to develop a robust statistical model to predict neonatal weight using a dataset from three hospitals. The focus is on analyzing various maternal and neonatal variables to assess their influence on newborn weight, with particular attention to the impact of maternal smoking.

Key Achievements

🔍 Data Mastery:

• Utilized a comprehensive dataset of 2500 newborns, enhancing understanding of key factors influencing neonatal weight.
• Executed meticulous data cleaning and verification to ensure the accuracy of the analysis.

🌟 Challenging Task Conquered:

• Developed predictive models incorporating multiple variables that significantly affect neonatal outcomes.
• Addressed complex statistical challenges, such as nonlinear relationships and interactions between variables.

💡 Innovative Approaches:

• Employed advanced inferential statistical methods to draw meaningful conclusions about neonatal weight influences.
• Implemented a multiple linear regression model, exploring beyond linear assumptions by investigating potential nonlinear effects and interactions.

Your Experience Journey

📊 Key Dataset Properties:

• The dataset includes variables such as mother's age, number of pregnancies, gestational age, and neonatal physical measurements.
• Captures key categorical data like maternal smoking, type of delivery, hospital ID, and sex of the newborn, providing a rich basis for multivariate analysis.

🔮 Your Impact:

• Significantly advanced the field of neonatal care by identifying critical maternal and neonatal factors that predict weight at birth.
• Enhanced decision-making tools for healthcare professionals, contributing to improved neonatal health strategies.

Explore My Code

🔗 GitHub Repository: Dive into the codebase to follow the journey of crafting a sophisticated statistical model in R. Discover how exploratory data analysis, hypothesis testing, and regression modeling come together to predict neonatal weight effectively. See how each analytical step contributes to a comprehensive understanding of the factors impacting newborn health.

Project Workflow

1. Data Import:

   • Ensured accurate import and handling of the neonati.csv dataset into the R environment.

2. Descriptive Analysis:

   • Thoroughly described dataset properties, focusing on variables critical to neonatal weight prediction.

3. Exploratory Data Analysis (EDA):

   • Utilized statistical indices and visual tools to uncover patterns and insights within the data.

4. Hypothesis Testing:

   • Tested hypotheses about differences in neonatal weights and lengths, across various subgroups including gender and hospital type.

5. Multivariate Analysis:

   • Developed and refined a multiple linear regression model, using rigorous criteria to select the most effective model.

6. Residual Analysis:

   • Performed detailed diagnostics to ensure the model’s reliability, identifying influential cases that could impact predictive performance.

7. Predictive Performance:

   • Evaluated the model's accuracy through real-world predictions, such as estimating the weight for a third pregnancy at the 39th week without ultrasound data.

Visual Representations and Further Analysis

• Model Visualization:
  • Created detailed graphical representations to make the statistical model's results accessible and understandable.