Build and deploy a stroke prediction model using R
ODIGIE BENISON BLESSING
2024-06-02
About Data Analysis Report
This RMarkdown file contains the report of the data analysis done for the project on building and deploying a stroke prediction model in R. It contains analysis such as data exploration, summary statistics and building the prediction models. The final report was completed on Sun Jun 2 20:35:02 2024.
Data Description:
According to the World Health Organization (WHO) stroke is the 2nd leading cause of death globally, responsible for approximately 11% of total deaths.
This data set is used to predict whether a patient is likely to get stroke based on the input parameters like gender, age, various diseases, and smoking status. Each row in the data provides relevant information about the patient.
Task One: Import data and data preprocessing
Load data and install packages
## vars n mean sd median trimmed mad min max
## gender* 1 5110 1.41 0.49 1.00 1.39 0.00 1.00 3.00
## age 2 5110 43.23 22.61 45.00 43.61 26.69 0.08 82.00
## hypertension 3 5110 0.10 0.30 0.00 0.00 0.00 0.00 1.00
## heart_disease 4 5110 0.05 0.23 0.00 0.00 0.00 0.00 1.00
## ever_married* 5 5110 1.66 0.48 2.00 1.70 0.00 1.00 2.00
## work_type* 6 5110 3.50 1.28 4.00 3.62 0.00 1.00 5.00
## Residence_type* 7 5110 1.51 0.50 2.00 1.51 0.00 1.00 2.00
## avg_glucose_level 8 5110 106.15 45.28 91.88 97.85 26.06 55.12 271.74
## bmi 9 4909 28.89 7.85 28.10 28.34 6.97 10.30 97.60
## smoking_status* 10 5110 2.59 1.09 2.00 2.61 1.48 1.00 4.00
## stroke 11 5110 0.05 0.22 0.00 0.00 0.00 0.00 1.00
## range skew kurtosis se
## gender* 2.00 0.35 -1.86 0.01
## age 81.92 -0.14 -0.99 0.32
## hypertension 1.00 2.71 5.37 0.00
## heart_disease 1.00 3.94 13.57 0.00
## ever_married* 1.00 -0.66 -1.57 0.01
## work_type* 4.00 -0.91 -0.49 0.02
## Residence_type* 1.00 -0.03 -2.00 0.01
## avg_glucose_level 216.62 1.57 1.68 0.63
## bmi 87.30 1.05 3.36 0.11
## smoking_status* 3.00 0.08 -1.35 0.02
## stroke 1.00 4.19 15.57 0.00Describe and explore the data
## vars n mean sd median trimmed mad min max
## gender* 1 5110 1.41 0.49 1.00 1.39 0.00 1.00 3.00
## age 2 5110 43.23 22.61 45.00 43.61 26.69 0.08 82.00
## hypertension 3 5110 0.10 0.30 0.00 0.00 0.00 0.00 1.00
## heart_disease 4 5110 0.05 0.23 0.00 0.00 0.00 0.00 1.00
## ever_married* 5 5110 1.66 0.48 2.00 1.70 0.00 1.00 2.00
## work_type* 6 5110 3.50 1.28 4.00 3.62 0.00 1.00 5.00
## Residence_type* 7 5110 1.51 0.50 2.00 1.51 0.00 1.00 2.00
## avg_glucose_level 8 5110 106.15 45.28 91.88 97.85 26.06 55.12 271.74
## bmi 9 4909 28.89 7.85 28.10 28.34 6.97 10.30 97.60
## smoking_status* 10 5110 2.59 1.09 2.00 2.61 1.48 1.00 4.00
## stroke 11 5110 0.05 0.22 0.00 0.00 0.00 0.00 1.00
## range skew kurtosis se
## gender* 2.00 0.35 -1.86 0.01
## age 81.92 -0.14 -0.99 0.32
## hypertension 1.00 2.71 5.37 0.00
## heart_disease 1.00 3.94 13.57 0.00
## ever_married* 1.00 -0.66 -1.57 0.01
## work_type* 4.00 -0.91 -0.49 0.02
## Residence_type* 1.00 -0.03 -2.00 0.01
## avg_glucose_level 216.62 1.57 1.68 0.63
## bmi 87.30 1.05 3.36 0.11
## smoking_status* 3.00 0.08 -1.35 0.02
## stroke 1.00 4.19 15.57 0.00Missing Value Computation and Data Preprocessing
## [1] "Are there missing values in the dataset? TRUE"## [1] "How many? 201"## [1] "What Proportion: 0.0035758761786159"## gender age hypertension heart_disease
## 0 0 0 0
## ever_married work_type Residence_type avg_glucose_level
## 0 0 0 0
## bmi smoking_status stroke
## 201 0 0
#Data spliting
## <Training/Testing/Total>
## <3832/1278/5110>## [1] "Balance Training dataset"## < table of extent 0 >## # A tibble: 3,832 × 11
## gender age hypertension heart_disease ever_married work_type
## <fct> <dbl> <dbl> <dbl> <fct> <fct>
## 1 Female -0.0904 -0.324 -0.240 Yes Private
## 2 Male -1.81 -0.324 -0.240 No children
## 3 Male -0.575 -0.324 -0.240 Yes Private
## 4 Female 1.01 -0.324 -0.240 Yes Private
## 5 Female -0.355 -0.324 -0.240 Yes Self-employed
## 6 Female 0.262 -0.324 -0.240 Yes Private
## 7 Male 0.703 -0.324 -0.240 Yes Govt_job
## 8 Male 0.130 3.08 -0.240 Yes Private
## 9 Female -1.02 -0.324 -0.240 No Private
## 10 Male -0.267 -0.324 -0.240 Yes Private
## # ℹ 3,822 more rows
## # ℹ 5 more variables: Residence_type <fct>, avg_glucose_level <dbl>, bmi <dbl>,
## # smoking_status <fct>, stroke <fct>Task Two: Build prediction models
## # A tibble: 20 × 7
## mtry .metric .estimator mean n std_err .config
## <int> <chr> <chr> <dbl> <int> <dbl> <chr>
## 1 5 accuracy binary 0.950 9 0.00292 Preprocessor1_Model01
## 2 5 roc_auc binary 0.825 9 0.0170 Preprocessor1_Model01
## 3 9 accuracy binary 0.950 9 0.00339 Preprocessor1_Model02
## 4 9 roc_auc binary 0.822 9 0.0158 Preprocessor1_Model02
## 5 2 accuracy binary 0.951 9 0.00310 Preprocessor1_Model03
## 6 2 roc_auc binary 0.833 9 0.0176 Preprocessor1_Model03
## 7 10 accuracy binary 0.950 9 0.00339 Preprocessor1_Model04
## 8 10 roc_auc binary 0.818 9 0.0173 Preprocessor1_Model04
## 9 7 accuracy binary 0.950 9 0.00296 Preprocessor1_Model05
## 10 7 roc_auc binary 0.821 9 0.0173 Preprocessor1_Model05
## 11 3 accuracy binary 0.950 9 0.00304 Preprocessor1_Model06
## 12 3 roc_auc binary 0.829 9 0.0178 Preprocessor1_Model06
## 13 1 accuracy binary 0.951 9 0.00310 Preprocessor1_Model07
## 14 1 roc_auc binary 0.834 9 0.0184 Preprocessor1_Model07
## 15 6 accuracy binary 0.950 9 0.00301 Preprocessor1_Model08
## 16 6 roc_auc binary 0.823 9 0.0189 Preprocessor1_Model08
## 17 4 accuracy binary 0.951 9 0.00273 Preprocessor1_Model09
## 18 4 roc_auc binary 0.827 9 0.0182 Preprocessor1_Model09
## 19 8 accuracy binary 0.950 9 0.00346 Preprocessor1_Model10
## 20 8 roc_auc binary 0.820 9 0.0171 Preprocessor1_Model10## [1] "------Rank result accounding to Wokflow set 1"Task Three: Evaluate and select prediction models
## # A tibble: 1 × 2
## mtry .config
## <int> <chr>
## 1 2 Preprocessor1_Model03## # Resampling results
## # Manual resampling
## # A tibble: 1 × 6
## splits id .metrics .notes .predictions .workflow
## <list> <chr> <list> <list> <list> <list>
## 1 <split [3832/1278]> train/test split <tibble> <tibble> <tibble> <workflow>## # A tibble: 3 × 4
## .metric .estimator .estimate .config
## <chr> <chr> <dbl> <chr>
## 1 accuracy binary 0.949 Preprocessor1_Model1
## 2 roc_auc binary 0.823 Preprocessor1_Model1
## 3 brier_class binary 0.0451 Preprocessor1_Model1## # A tibble: 1,278 × 7
## .pred_class .pred_heart_disease .pred_Normal id .row stroke .config
## <fct> <dbl> <dbl> <chr> <int> <fct> <chr>
## 1 Normal 0.0423 0.958 train/test… 1 heart… Prepro…
## 2 Normal 0.262 0.738 train/test… 3 heart… Prepro…
## 3 Normal 0.0755 0.925 train/test… 11 heart… Prepro…
## 4 Normal 0.200 0.800 train/test… 12 heart… Prepro…
## 5 Normal 0.229 0.771 train/test… 13 heart… Prepro…
## 6 Normal 0.199 0.801 train/test… 14 heart… Prepro…
## 7 Normal 0.0768 0.923 train/test… 15 heart… Prepro…
## 8 Normal 0.0341 0.966 train/test… 21 heart… Prepro…
## 9 Normal 0.162 0.838 train/test… 33 heart… Prepro…
## 10 Normal 0.130 0.870 train/test… 34 heart… Prepro…
## # ℹ 1,268 more rows## Truth
## Prediction heart_disease Normal
## heart_disease 0 0
## Normal 65 1213## # A tibble: 1,278 × 7
## .pred_class .pred_heart_disease .pred_Normal id .row stroke .config
## <fct> <dbl> <dbl> <chr> <int> <fct> <chr>
## 1 Normal 0.0423 0.958 train/test… 1 heart… Prepro…
## 2 Normal 0.262 0.738 train/test… 3 heart… Prepro…
## 3 Normal 0.0755 0.925 train/test… 11 heart… Prepro…
## 4 Normal 0.200 0.800 train/test… 12 heart… Prepro…
## 5 Normal 0.229 0.771 train/test… 13 heart… Prepro…
## 6 Normal 0.199 0.801 train/test… 14 heart… Prepro…
## 7 Normal 0.0768 0.923 train/test… 15 heart… Prepro…
## 8 Normal 0.0341 0.966 train/test… 21 heart… Prepro…
## 9 Normal 0.162 0.838 train/test… 33 heart… Prepro…
## 10 Normal 0.130 0.870 train/test… 34 heart… Prepro…
## # ℹ 1,268 more rows
## [[1]]
## # A tibble: 1,278 × 6
## .pred_heart_disease .pred_Normal .row .pred_class stroke .config
## <dbl> <dbl> <int> <fct> <fct> <chr>
## 1 0.0423 0.958 1 Normal heart_disease Preprocesso…
## 2 0.262 0.738 3 Normal heart_disease Preprocesso…
## 3 0.0755 0.925 11 Normal heart_disease Preprocesso…
## 4 0.200 0.800 12 Normal heart_disease Preprocesso…
## 5 0.229 0.771 13 Normal heart_disease Preprocesso…
## 6 0.199 0.801 14 Normal heart_disease Preprocesso…
## 7 0.0768 0.923 15 Normal heart_disease Preprocesso…
## 8 0.0341 0.966 21 Normal heart_disease Preprocesso…
## 9 0.162 0.838 33 Normal heart_disease Preprocesso…
## 10 0.130 0.870 34 Normal heart_disease Preprocesso…
## # ℹ 1,268 more rows## ══ Workflow [trained] ══════════════════════════════════════════════════════════
## Preprocessor: Recipe
## Model: rand_forest()
##
## ── Preprocessor ────────────────────────────────────────────────────────────────
## 3 Recipe Steps
##
## • step_normalize()
## • step_zv()
## • step_corr()
##
## ── Model ───────────────────────────────────────────────────────────────────────
## Ranger result
##
## Call:
## ranger::ranger(x = maybe_data_frame(x), y = y, mtry = min_cols(~2L, x), importance = ~"impurity", num.threads = 1, verbose = FALSE, seed = sample.int(10^5, 1), probability = TRUE)
##
## Type: Probability estimation
## Number of trees: 500
## Sample size: 5110
## Number of independent variables: 10
## Mtry: 2
## Target node size: 10
## Variable importance mode: impurity
## Splitrule: gini
## OOB prediction error (Brier s.): 0.04332416Task Four: Deploy the prediction model
## # A tibble: 1 × 10
## gender age hypertension heart_disease ever_married work_type Residence_type
## <chr> <dbl> <dbl> <dbl> <chr> <chr> <chr>
## 1 Female 47 0 0 Yes Private Urban
## # ℹ 3 more variables: avg_glucose_level <dbl>, bmi <dbl>, smoking_status <chr>## # A tibble: 1 × 1
## .pred_class
## <fct>
## 1 NormalTask Five: Findings and Conclusions
#Variable Importance
ranger_obj <- pull_workflow_fit(final_model)$fit
ranger_obj## Ranger result
##
## Call:
## ranger::ranger(x = maybe_data_frame(x), y = y, mtry = min_cols(~2L, x), importance = ~"impurity", num.threads = 1, verbose = FALSE, seed = sample.int(10^5, 1), probability = TRUE)
##
## Type: Probability estimation
## Number of trees: 500
## Sample size: 5110
## Number of independent variables: 10
## Mtry: 2
## Target node size: 10
## Variable importance mode: impurity
## Splitrule: gini
## OOB prediction error (Brier s.): 0.04332416#Display VAriable Importance
sort(ranger_obj$variable.importance)## gender Residence_type ever_married hypertension
## 5.009274 5.363850 5.942271 7.099011
## heart_disease work_type smoking_status bmi
## 7.877685 9.870189 11.611983 43.775349
## avg_glucose_level age
## 56.908398 59.367519