Scientific Calculations with Julia/Diagnosing Heart Disease

Scientific Calculations with Julia Diagnosing Heart Disease Electrical Circuits SIR Model Loading Libraries First, you need to install and use the following libraries: using CSV using DataFrames using GLM using StatsBase:mean Part 1: Data The "heart.csv" dataset, downloadable from Kaggle, contains 1,025 observations, each containing 14 heart-related variables. These variables are as follows: age: the person's age in years sex: the person's gender (1 = male, 0 = female) cp: chest pain experienced (1: typical angina, 2: atypical angina, 3: non-anginal pain, 4: asymptomatic) trestbps: the person's resting blood pressure (mm/Hg at the time of hospital admission) chol: cholesterol measurement in mg/dL fbs: the person's fasting blood glucose (> 120 mg/dL, 1 = true; 0 = false) restecg: resting electrocardiogram (0 = normal, 1 = with ST-T wave abnormality, 2 = probable or definite left ventricular hypertrophy according to Estes criteria) thalach: the person's maximum heart rate exang: exercise-induced angina (1 = yes; 0 = (no) oldpeak: exercise-induced ST depression ('ST' refers to the positions on the ECG diagram) slop: ST segment slope (value 1: rising, value 2: flat, value 3: falling) ca: number of major vessels (0-3) tal: blood disorder called thalassemia (3 = normal; 6 = fixed defect; 7 = reversible defect) target: heart disease (0 = no, 1 = yes) Loading data: data = CSV.read("heart.csv",DataFrame) Part 2: Research Question We want to predict the possible presence of heart disease in a patient by predicting the target variable based on the 13 aforementioned variables. Part 3: Modeling and Prediction The dataset is divided into two parts: train_data, consisting of 80% of the observations on which the logistic regression model is trained, and test_data, consisting of the remaining 20% ​​on which the model is tested and its accuracy is calculated: train_ratio = 0.8 n = nrow(data) n_train = Int(floor(train_ratio * n)); train_data = data[1:n_train, :] test_data = data[(n_train + 1):end, :] The logistic regression model is created on the training set, predicting the target variable based on the remaining 13 variables: fm = @formula(target ~ age + sex + cp + trestbps + chol + fbs + restecg + thalach + exang + oldpeak + slope + ca + thal) logit = glm(fm, train_data, Binomial(), ProbitLink()) The prediction vector is constructed with the predictions made by the model on the testing set: prediction = GLM.predict(logit, test_data) 205-element Vector{Union{Missing, Float64}}: 0.017732763641414958 0.06652842473470792 0.04289852866536204 0.5470639026550284 0.006095278913460305 0.9955768152608755 0.9703221524362567 0.5142923144983738 0.9053830876384388 0.0023961543663923637 0.6965675386130677 0.04394901761879894 0.5745052037713632 ⋮ 0.005816355358123194 0.9984324380351746 0.0003433323966801158 0.7506524853801526 0.000691630070259472 0.6261769123744736 0.7526861512703369 0.7105401597352763 0.0033310752534602382 0.05287226311321894 0.949991107720036 0.06892902969480652 A vector of 1s and 0s is constructed based on the probabilities contained in the prediction vector. If the probability is greater than 50%, it takes the value 1, otherwise it is 0. Then the accuracy is calculated, which is 80.48%: prediction_class = [if x < 0.5 0 else 1 end for x in prediction]; accuracy = mean(test_data.target.==prediction_class) 0.8048780487804879 Therefore, logistic regression is a good predictive model for assessing the presence of heart disease.