clinical-text-mining_R_SCRIPT#

🏥 Medical Phenotype Extraction from Doctor's Notes 🩺

📜 Overview

This R script extracts structured phenotype data from unstructured doctor's notes.
It cleans, standardizes, maps diagnoses to ICD-10 codes, applies one-hot encoding,
and exports a ready-to-use phenotype matrix for machine learning & statistical analysis.

🔬 Features

✅ Parses doctor’s notes into structured data using regex & NLP
✅ Handles missing values & normalizes blood pressure, weight, age
✅ Maps diagnoses to ICD-10 codes for standardization
✅ One-hot encodes categorical data (diagnosis & meds) for ML
✅ Saves phenotype_matrix.csv for database integration & research

---

⚙️ Installation & Dependencies

install.packages(c("dplyr", "tidyr", "stringr"))

🚀 Usage
Prepare your raw doctor’s notes in a structured text file.
Run the script to extract structured data:
r
Copy
Edit
source("generate_phenotype_matrix.R")
Upload the phenotype_matrix.csv to your lab’s database.
📂 Output Example
sample_id	age	weight_kg	systolic	diastolic	diagnosis_Hypertension	diagnosis_Diabetes	diagnosis_Asthma	diagnosis_Cardiovascular_Disease	med_Lisinopril	med_Metformin	med_Albuterol	med_Atorvastatin
S001	56	81	140	90	1	0	0	0	1	0	0	0
🏥 Database Integration
If using SQL, run:


library(DBI)
con <- dbConnect(RSQLite::SQLite(), dbname = "lab_database.sqlite")
dbWriteTable(con, "phenotype_data", read.csv("phenotype_matrix.csv"), overwrite = TRUE)
dbDisconnect(con)