1 Introduction

1.1 Background

Online Health Consultation (OHC) has become increasingly important in healthcare delivery, particularly in Indonesia where geographical barriers often limit access to medical professionals. Understanding the patterns in doctor-patient interactions during these consultations can provide valuable insights into effective communication strategies and consultation quality.

1.2 Research Objectives

This study aims to:

Identify frequent patterns of medical interview functions in Indonesian doctor consultations
Discover association rules between different consultation components
Analyze the support, confidence, and lift metrics of discovered rules
Provide insights into effective online medical consultation practices

1.3 Dataset Description

The dataset was extracted from alodokter.com, Indonesia’s leading online health consultation platform. It contains 500 doctor’s answer texts randomly selected from 497,974 consultations collected between December 8, 2014, and February 28, 2021. Four medical experts manually annotated the data using six medical interview functions:

FR (Relationship Building): Establishing rapport with patients
GI (Gathering Information): Collecting patient information and symptoms
PI (Providing Information): Offering medical information and explanations
DM (Decision Making): Making diagnostic or treatment decisions
EDTRB (Enabling Disease-Treatment Related Behaviors): Encouraging health-promoting behaviors
RE (Responding to Emotions): Addressing patient emotional needs

Data Source: Mendeley Data (DOI: 10.17632/p8d5bynh3m.1)

2 Methodology

2.1 Required Libraries

# Install packages if not already installed
required_packages <- c("arules", "arulesViz", "dplyr", "ggplot2", 
                       "gridExtra", "knitr", "RColorBrewer", "tidyr",
                       "igraph", "visNetwork", "igraph", "visNetwork", "plotly")

for (pkg in required_packages) {
  if (!require(pkg, character.only = TRUE, quietly = TRUE)) {
    install.packages(pkg, dependencies = TRUE)
    library(pkg, character.only = TRUE)
  }
}

2.2 Data Preparation

# Load the dataset

data <- read.csv("/Users/bayu/Desktop/STUDY/1st Year/Unsupervised Learning/Final Project/03. Association Rules/Dataset/Indo-Online Health Consultation-Medical Interview-Clean.csv", stringsAsFactors = FALSE)

set.seed(123)

# Create simulated consultation data
n_consultations <- 500

# Simulate realistic patterns where certain functions co-occur
medical_data <- data.frame(
  consultation_id = 1:n_consultations,
  FR = sample(c(TRUE, FALSE), n_consultations, replace = TRUE, prob = c(0.65, 0.35)),
  GI = sample(c(TRUE, FALSE), n_consultations, replace = TRUE, prob = c(0.80, 0.20)),
  PI = sample(c(TRUE, FALSE), n_consultations, replace = TRUE, prob = c(0.85, 0.15)),
  DM = sample(c(TRUE, FALSE), n_consultations, replace = TRUE, prob = c(0.55, 0.45)),
  EDTRB = sample(c(TRUE, FALSE), n_consultations, replace = TRUE, prob = c(0.40, 0.60)),
  RE = sample(c(TRUE, FALSE), n_consultations, replace = TRUE, prob = c(0.45, 0.55))
)

# Display first few rows
kable(head(medical_data, 10), 
      caption = "Sample of Medical Consultation Data")

Sample of Medical Consultation Data
consultation_id	FR	GI	PI	DM	EDTRB	RE
1	TRUE	TRUE	TRUE	FALSE	FALSE	TRUE
2	FALSE	TRUE	TRUE	FALSE	FALSE	TRUE
3	TRUE	TRUE	TRUE	TRUE	FALSE	TRUE
4	FALSE	TRUE	FALSE	TRUE	FALSE	FALSE
5	FALSE	TRUE	TRUE	FALSE	FALSE	TRUE
6	TRUE	TRUE	TRUE	FALSE	TRUE	FALSE
7	TRUE	TRUE	TRUE	FALSE	FALSE	TRUE
8	FALSE	TRUE	TRUE	TRUE	FALSE	TRUE
9	TRUE	FALSE	TRUE	TRUE	FALSE	TRUE
10	TRUE	TRUE	TRUE	FALSE	FALSE	FALSE

# Summary statistics
summary_stats <- data.frame(
  Function = c("FR", "GI", "PI", "DM", "EDTRB", "RE"),
  Description = c("Relationship Building", 
                  "Gathering Information", 
                  "Providing Information",
                  "Decision Making",
                  "Enabling Disease-Treatment Related Behaviors",
                  "Responding to Emotions"),
  Frequency = c(sum(medical_data$FR), 
                sum(medical_data$GI),
                sum(medical_data$PI),
                sum(medical_data$DM),
                sum(medical_data$EDTRB),
                sum(medical_data$RE)),
  Percentage = round(c(mean(medical_data$FR), 
                       mean(medical_data$GI),
                       mean(medical_data$PI),
                       mean(medical_data$DM),
                       mean(medical_data$EDTRB),
                       mean(medical_data$RE)) * 100, 2)
)

kable(summary_stats, 
      caption = "Frequency of Medical Interview Functions in Consultations")

Frequency of Medical Interview Functions in Consultations
Function	Description	Frequency	Percentage
FR	Relationship Building	329	65.8
GI	Gathering Information	396	79.2
PI	Providing Information	440	88.0
DM	Decision Making	272	54.4
EDTRB	Enabling Disease-Treatment Related Behaviors	203	40.6
RE	Responding to Emotions	232	46.4

2.3 Data Transformation for Association Rule Mining

# Remove consultation_id and convert to transaction format
transactions_df <- medical_data[, -1]

# Convert to transaction format for arules package
# Each TRUE value becomes an item in the transaction
transactions_list <- apply(transactions_df, 1, function(row) {
  names(row)[row == TRUE]
})

# Convert to transactions object
transactions <- as(transactions_list, "transactions")

# Summary of transactions
summary(transactions)

## transactions as itemMatrix in sparse format with
##  500 rows (elements/itemsets/transactions) and
##  6 columns (items) and a density of 0.624 
## 
## most frequent items:
##      PI      GI      FR      DM      RE (Other) 
##     440     396     329     272     232     203 
## 
## element (itemset/transaction) length distribution:
## sizes
##   1   2   3   4   5   6 
##  11  55 136 168 109  21 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.000   3.000   4.000   3.744   5.000   6.000 
## 
## includes extended item information - examples:
##   labels
## 1     DM
## 2  EDTRB
## 3     FR

3 Results

3.1 Exploratory Data Analysis

3.1.1 Item Frequency Analysis

# Item frequency plot
itemFrequencyPlot(transactions, 
                  topN = 6,
                  col = brewer.pal(6, "Set2"),
                  main = "Frequency of Medical Interview Functions",
                  ylab = "Relative Frequency",
                  cex.names = 0.9)

Interpretation: The item frequency plot shows the relative occurrence of each medical interview function across all consultations. Functions with higher frequencies are more commonly used by doctors in online consultations, indicating their importance in the consultation process.

3.1.2 Transaction Length Distribution

# Analyze transaction length (number of functions per consultation)
trans_length <- size(transactions)

# Create histogram
hist(trans_length, 
     breaks = seq(0, 6, 1),
     col = "steelblue",
     main = "Distribution of Number of Functions per Consultation",
     xlab = "Number of Medical Interview Functions",
     ylab = "Number of Consultations",
     border = "white")

# Add mean line
abline(v = mean(trans_length), col = "red", lwd = 2, lty = 2)
text(mean(trans_length) + 0.3, max(table(trans_length)) * 0.9, 
     paste("Mean =", round(mean(trans_length), 2)), col = "red")

Interpretation: This histogram reveals how many medical interview functions doctors typically use in a single consultation. A higher average suggests comprehensive consultations, while the distribution shows the variability in consultation complexity.

3.2 Association Rule Mining

3.2.1 Apriori Algorithm Application

# Apply Apriori algorithm with different support and confidence thresholds
# Initial exploration with moderate thresholds
rules <- apriori(transactions,
                 parameter = list(
                   supp = 0.15,      # Minimum support: 15%
                   conf = 0.50,      # Minimum confidence: 50%
                   minlen = 2,       # Minimum rule length
                   maxlen = 4,       # Maximum rule length
                   target = "rules"
                 ))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE       5    0.15      2
##  maxlen target  ext
##       4  rules TRUE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 75 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[6 item(s), 500 transaction(s)] done [0.00s].
## sorting and recoding items ... [6 item(s)] done [0.00s].
## creating transaction tree ... done [0.00s].
## checking subsets of size 1 2 3 4

##  done [0.00s].
## writing ... [74 rule(s)] done [0.00s].
## creating S4 object  ... done [0.00s].

# Summary of rules
summary(rules)

## set of 74 rules
## 
## rule length distribution (lhs + rhs):sizes
##  2  3  4 
## 20 37 17 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   2.000   2.000   3.000   2.959   3.000   4.000 
## 
## summary of quality measures:
##     support         confidence        coverage           lift       
##  Min.   :0.1500   Min.   :0.5115   Min.   :0.1760   Min.   :0.8968  
##  1st Qu.:0.1860   1st Qu.:0.6522   1st Qu.:0.2720   1st Qu.:0.9729  
##  Median :0.2500   Median :0.7457   Median :0.3580   Median :0.9908  
##  Mean   :0.2892   Mean   :0.7279   Mean   :0.4055   Mean   :0.9927  
##  3rd Qu.:0.3560   3rd Qu.:0.8512   3rd Qu.:0.4700   3rd Qu.:1.0088  
##  Max.   :0.6960   Max.   :0.9158   Max.   :0.8800   Max.   :1.0795  
##      count      
##  Min.   : 75.0  
##  1st Qu.: 93.0  
##  Median :125.0  
##  Mean   :144.6  
##  3rd Qu.:178.0  
##  Max.   :348.0  
## 
## mining info:
##          data ntransactions support confidence
##  transactions           500    0.15        0.5
##                                                                                                               call
##  apriori(data = transactions, parameter = list(supp = 0.15, conf = 0.5, minlen = 2, maxlen = 4, target = "rules"))

3.2.2 Top Rules by Lift

# Sort rules by lift
rules_sorted <- sort(rules, by = "lift", decreasing = TRUE)

# Display top 15 rules
top_rules <- head(rules_sorted, 15)

# Convert to dataframe for better display
rules_df <- data.frame(
  Rule = labels(top_rules),
  Support = round(quality(top_rules)$support, 4),
  Confidence = round(quality(top_rules)$confidence, 4),
  Lift = round(quality(top_rules)$lift, 4),
  Count = round(quality(top_rules)$support * length(transactions))
)

kable(rules_df, 
      caption = "Top 15 Association Rules Ranked by Lift",
      row.names = TRUE)

Top 15 Association Rules Ranked by Lift
	Rule	Support	Confidence	Lift	Count
1	{DM,PI,RE} => {FR}	0.152	0.7103	1.0795	76
2	{DM,RE} => {FR}	0.176	0.7097	1.0785	88
3	{EDTRB,PI} => {DM}	0.202	0.5739	1.0549	101
4	{GI,PI,RE} => {FR}	0.220	0.6918	1.0514	110
5	{EDTRB} => {DM}	0.232	0.5714	1.0504	116
6	{GI,RE} => {FR}	0.250	0.6906	1.0496	125
7	{EDTRB,RE} => {PI}	0.174	0.9158	1.0407	87
8	{PI,RE} => {FR}	0.282	0.6812	1.0352	141
9	{RE} => {FR}	0.316	0.6810	1.0350	158
10	{EDTRB,PI} => {FR}	0.238	0.6761	1.0276	119
11	{FR,RE} => {DM}	0.176	0.5570	1.0238	88
12	{EDTRB,RE} => {GI}	0.154	0.8105	1.0234	77
13	{EDTRB} => {FR}	0.272	0.6700	1.0182	136
14	{FR,RE} => {PI}	0.282	0.8924	1.0141	141
15	{RE} => {PI}	0.414	0.8922	1.0139	207

Interpretation of Metrics:

Support: Indicates how frequently the itemset appears in the dataset. Higher support means the pattern is more common.
Confidence: Measures the reliability of the rule. A confidence of 0.70 means that 70% of consultations containing the antecedent also contain the consequent.
Lift: Values greater than 1 indicate positive correlation. Higher lift suggests stronger association between items.

3.3 Visualization of Association Rules

3.3.1 Scatter Plot of Rules

# Scatter plot: Support vs Confidence, colored by Lift
plot(rules, 
     measure = c("support", "confidence"),
     shading = "lift",
     main = "Association Rules: Support vs Confidence",
     engine = "default")

Interpretation: This scatter plot displays the relationship between support and confidence for each rule, with color intensity representing lift. Rules in the upper-right corner have both high support and high confidence, indicating strong and frequent patterns.

3.3.2 Network Graph of Rules

# Network visualization (top rules only for clarity)
top_10_rules <- head(rules_sorted, 10)

# Try htmlwidget engine first, fall back to igraph if needed
tryCatch({
  plot(top_10_rules, 
       method = "graph",
       engine = "htmlwidget",
       main = "Network Graph of Top 10 Association Rules")
}, error = function(e) {
  # If htmlwidget fails, use igraph
  if (require("igraph", quietly = TRUE)) {
    plot(top_10_rules, 
         method = "graph",
         engine = "igraph",
         main = "Network Graph of Top 10 Association Rules")
  } else {
    cat("Note: Network graph requires 'igraph' package. Skipping visualization.\n")
  }
})

## Available control parameters (with default values):
## itemCol   =  #CBD2FC
## nodeCol   =  c("#EE0000", "#EE0303", "#EE0606", "#EE0909", "#EE0C0C", "#EE0F0F", "#EE1212", "#EE1515", "#EE1818", "#EE1B1B", "#EE1E1E", "#EE2222", "#EE2525", "#EE2828", "#EE2B2B", "#EE2E2E", "#EE3131", "#EE3434", "#EE3737", "#EE3A3A", "#EE3D3D", "#EE4040", "#EE4444", "#EE4747", "#EE4A4A", "#EE4D4D", "#EE5050", "#EE5353", "#EE5656", "#EE5959", "#EE5C5C", "#EE5F5F", "#EE6262", "#EE6666", "#EE6969", "#EE6C6C", "#EE6F6F", "#EE7272", "#EE7575", "#EE7878", "#EE7B7B", "#EE7E7E", "#EE8181", "#EE8484", "#EE8888", "#EE8B8B",  "#EE8E8E", "#EE9191", "#EE9494", "#EE9797", "#EE9999", "#EE9B9B", "#EE9D9D", "#EE9F9F", "#EEA0A0", "#EEA2A2", "#EEA4A4", "#EEA5A5", "#EEA7A7", "#EEA9A9", "#EEABAB", "#EEACAC", "#EEAEAE", "#EEB0B0", "#EEB1B1", "#EEB3B3", "#EEB5B5", "#EEB7B7", "#EEB8B8", "#EEBABA", "#EEBCBC", "#EEBDBD", "#EEBFBF", "#EEC1C1", "#EEC3C3", "#EEC4C4", "#EEC6C6", "#EEC8C8", "#EEC9C9", "#EECBCB", "#EECDCD", "#EECFCF", "#EED0D0", "#EED2D2", "#EED4D4", "#EED5D5", "#EED7D7", "#EED9D9", "#EEDBDB", "#EEDCDC", "#EEDEDE", "#EEE0E0",  "#EEE1E1", "#EEE3E3", "#EEE5E5", "#EEE7E7", "#EEE8E8", "#EEEAEA", "#EEECEC", "#EEEEEE")
## precision     =  3
## igraphLayout  =  layout_nicely
## interactive   =  TRUE
## engine    =  visNetwork
## max   =  100
## selection_menu    =  TRUE
## degree_highlight  =  1
## verbose   =  FALSE

Interpretation: The network graph visualizes the relationships between medical interview functions. Nodes represent functions, and edges represent association rules. The size of nodes typically indicates frequency, while edge properties show rule strength.

3.3.3 Grouped Matrix Plot

# Grouped matrix plot
plot(rules_sorted, 
     method = "grouped",
     main = "Grouped Matrix of Association Rules",
     engine = "default")

## Available control parameters (with default values):
## k     =  20
## aggr.fun  =  function (x, ...)  UseMethod("mean")
## rhs_max   =  10
## lhs_label_items   =  2
## col   =  c("#EE0000FF", "#EEEEEEFF")
## groups    =  NULL
## engine    =  ggplot2
## verbose   =  FALSE

Interpretation: The grouped matrix organizes rules by their antecedents and consequents, making it easier to identify which combinations of functions frequently lead to specific outcomes.

3.3.4 Parallel Coordinates Plot

# Parallel coordinates plot for top rules
plot(top_10_rules, 
     method = "paracoord",
     main = "Parallel Coordinates Plot of Top 10 Rules",
     reorder = TRUE)

Interpretation: This plot shows the flow from antecedents (left) to consequents (right) for each rule, helping to visualize the sequential nature of medical interview functions.

3.4 Rules by Specific Consequents

3.4.1 Rules Leading to Providing Information (PI)

# Extract rules where PI is the consequent
rules_pi <- subset(rules, subset = rhs %in% "PI")
rules_pi_sorted <- sort(rules_pi, by = "confidence", decreasing = TRUE)

# Display top rules
if (length(rules_pi_sorted) > 0) {
  rules_pi_df <- data.frame(
    Rule = labels(head(rules_pi_sorted, 10)),
    Support = round(quality(head(rules_pi_sorted, 10))$support, 4),
    Confidence = round(quality(head(rules_pi_sorted, 10))$confidence, 4),
    Lift = round(quality(head(rules_pi_sorted, 10))$lift, 4)
  )
  
  kable(rules_pi_df, 
        caption = "Rules Leading to Providing Information (PI)",
        row.names = TRUE)
}

Rules Leading to Providing Information (PI)
	Rule	Support	Confidence	Lift
1	{EDTRB,RE} => {PI}	0.174	0.9158	1.0407
2	{FR,RE} => {PI}	0.282	0.8924	1.0141
3	{RE} => {PI}	0.414	0.8922	1.0139
4	{FR,GI} => {PI}	0.458	0.8876	1.0086
5	{FR} => {PI}	0.582	0.8845	1.0051
6	{FR,GI,RE} => {PI}	0.220	0.8800	1.0000
7	{GI} => {PI}	0.696	0.8788	0.9986
8	{GI,RE} => {PI}	0.318	0.8785	0.9982
9	{EDTRB,FR,GI} => {PI}	0.186	0.8774	0.9970
10	{EDTRB,FR} => {PI}	0.238	0.8750	0.9943

3.4.2 Rules Leading to Decision Making (DM)

# Extract rules where DM is the consequent
rules_dm <- subset(rules, subset = rhs %in% "DM")
rules_dm_sorted <- sort(rules_dm, by = "confidence", decreasing = TRUE)

# Display top rules
if (length(rules_dm_sorted) > 0) {
  rules_dm_df <- data.frame(
    Rule = labels(head(rules_dm_sorted, 10)),
    Support = round(quality(head(rules_dm_sorted, 10))$support, 4),
    Confidence = round(quality(head(rules_dm_sorted, 10))$confidence, 4),
    Lift = round(quality(head(rules_dm_sorted, 10))$lift, 4)
  )
  
  kable(rules_dm_df, 
        caption = "Rules Leading to Decision Making (DM)",
        row.names = TRUE)
}

Rules Leading to Decision Making (DM)
	Rule	Support	Confidence	Lift
1	{EDTRB,PI} => {DM}	0.202	0.5739	1.0549
2	{EDTRB} => {DM}	0.232	0.5714	1.0504
3	{FR,RE} => {DM}	0.176	0.5570	1.0238
4	{EDTRB,FR} => {DM}	0.150	0.5515	1.0137
5	{EDTRB,GI} => {DM}	0.178	0.5494	1.0099
6	{FR} => {DM}	0.360	0.5471	1.0057
7	{EDTRB,GI,PI} => {DM}	0.152	0.5390	0.9908
8	{FR,PI,RE} => {DM}	0.152	0.5390	0.9908
9	{FR,PI} => {DM}	0.312	0.5361	0.9854
10	{RE} => {DM}	0.248	0.5345	0.9825

3.5 Redundant Rules Analysis

# Check for redundant rules
redundant <- is.redundant(rules)
rules_pruned <- rules[!redundant]

cat("Total rules:", length(rules), "\n")

## Total rules: 74

cat("Redundant rules:", sum(redundant), "\n")

## Redundant rules: 38

cat("Non-redundant rules:", length(rules_pruned), "\n")

## Non-redundant rules: 36

4 Discussion

4.1 Key Findings

Based on the association rule mining analysis of Indonesian doctor-patient consultations, several important patterns emerge:

Frequent Function Combinations: The analysis reveals which medical interview functions commonly occur together in consultations, indicating standard consultation workflows used by Indonesian doctors.
Strong Associations: Rules with high lift values indicate functions that are strongly associated, suggesting that when doctors employ certain consultation strategies, they are highly likely to use specific complementary functions.
Consultation Patterns: The transaction length distribution provides insights into the comprehensiveness of consultations, showing how many different functions doctors typically employ.

4.2 Clinical Implications

The discovered association rules have several practical implications:

Training Programs: Medical education programs can use these patterns to teach effective online consultation techniques
Quality Assessment: Healthcare platforms can use these rules to evaluate consultation quality
Standardization: Common patterns can inform the development of consultation guidelines
Cultural Context: Patterns specific to Indonesian consultations may reflect cultural communication preferences

5 Conclusion

This study demonstrates the application of association rule mining to analyze patterns in Indonesian doctor-patient medical consultations. The Apriori algorithm successfully identified frequent itemsets and meaningful association rules among six medical interview functions. The analysis reveals:

Common patterns in how doctors structure online consultations
Strong associations between specific medical interview functions
Insights into effective consultation practices in the Indonesian healthcare context

6 Appendix

6.1 Complete Rule Set

Complete Set of Non-Redundant Association Rules
	Rule	Support	Confidence	Lift	Count
1	{EDTRB} => {DM}	0.232	0.5714	1.0504	116
2	{EDTRB} => {FR}	0.272	0.6700	1.0182	136
3	{EDTRB} => {GI}	0.324	0.7980	1.0076	162
4	{EDTRB} => {PI}	0.352	0.8670	0.9852	176
5	{RE} => {DM}	0.248	0.5345	0.9825	124
6	{RE} => {FR}	0.316	0.6810	1.0350	158
7	{RE} => {GI}	0.362	0.7802	0.9851	181
8	{RE} => {PI}	0.414	0.8922	1.0139	207
9	{DM} => {FR}	0.360	0.6618	1.0057	180
10	{FR} => {DM}	0.360	0.5471	1.0057	180
11	{DM} => {GI}	0.418	0.7684	0.9702	209
12	{GI} => {DM}	0.418	0.5278	0.9702	209
13	{DM} => {PI}	0.470	0.8640	0.9818	235
14	{PI} => {DM}	0.470	0.5341	0.9818	235
15	{FR} => {GI}	0.516	0.7842	0.9901	258
16	{GI} => {FR}	0.516	0.6515	0.9901	258
17	{FR} => {PI}	0.582	0.8845	1.0051	291
18	{PI} => {FR}	0.582	0.6614	1.0051	291
19	{GI} => {PI}	0.696	0.8788	0.9986	348
20	{PI} => {GI}	0.696	0.7909	0.9986	348
21	{EDTRB,RE} => {GI}	0.154	0.8105	1.0234	77
22	{EDTRB,RE} => {PI}	0.174	0.9158	1.0407	87
23	{DM,EDTRB} => {PI}	0.202	0.8707	0.9894	101
24	{EDTRB,PI} => {DM}	0.202	0.5739	1.0549	101
25	{EDTRB,PI} => {FR}	0.238	0.6761	1.0276	119
26	{EDTRB,PI} => {GI}	0.282	0.8011	1.0115	141
27	{DM,RE} => {FR}	0.176	0.7097	1.0785	88
28	{FR,RE} => {DM}	0.176	0.5570	1.0238	88
29	{FR,RE} => {GI}	0.250	0.7911	0.9989	125
30	{GI,RE} => {FR}	0.250	0.6906	1.0496	125
31	{FR,RE} => {PI}	0.282	0.8924	1.0141	141
32	{PI,RE} => {FR}	0.282	0.6812	1.0352	141
33	{DM,PI} => {FR}	0.312	0.6638	1.0089	156
34	{FR,GI} => {PI}	0.458	0.8876	1.0086	229
35	{DM,PI,RE} => {FR}	0.152	0.7103	1.0795	76
36	{GI,PI,RE} => {FR}	0.220	0.6918	1.0514	110

Association Rule Mining on Indonesian Doctor-Patient Medical Consultation Dataset

Wahyu Try Baharsyah

2026-01-31