Add File

2025-09-04 14:08:43 +08:00
parent e0ac4405fa
commit 0e94ad39c7
1 changed files with 112 additions and 0 deletions
--- a/src/main/java/org/dromara/easyai/bayesian/NativeBayesian.java
+++ b/src/main/java/org/dromara/easyai/bayesian/NativeBayesian.java
@@ -0,0 +1,112 @@
+package org.dromara.easyai.bayesian;
+
+import org.dromara.easyai.randomForest.DataTable;
+
+import java.lang.reflect.Method;
+import java.util.*;
+
+public class NativeBayesian {
+    private DataTable dataTable;
+    // 存储每个类别的先验概率
+    private final Map<Integer, Double> priorProbabilities;
+    // 存储每个特征在每个类别下的条件概率
+    private final Map<String, Map<Integer, Map<Integer, Double>>> conditionalProbabilities;
+
+    public NativeBayesian() {
+        priorProbabilities = new HashMap<>();
+        conditionalProbabilities = new HashMap<>();
+    }
+
+    public NativeBayesian(DataTable dataTable) {
+        this();
+        this.dataTable = dataTable;
+    }
+
+    public DataTable getDataTable() {
+        return dataTable;
+    }
+
+    public void setDataTable(DataTable dataTable) {
+        this.dataTable = dataTable;
+    }
+
+    public int classify(Object object) {
+        // 存储每个类别的后验概率
+        Map<Integer, Double> posteriorProbabilities = new HashMap<>();
+        try {
+            // 获取所有可能的类别
+            List<Integer> classValues = dataTable.getTable().get(dataTable.getKey());
+            Set<Integer> uniqueClasses = new HashSet<>(classValues);
+
+            // 遍历每个类别
+            for (int classValue : uniqueClasses) {
+                double posteriorProb = priorProbabilities.get(classValue);
+                // 遍历每个特征
+                for (String feature : dataTable.getKeyType()) {
+                    if (!feature.equals(dataTable.getKey())) {
+                        String methodName = "get" + feature.substring(0, 1).toUpperCase() + feature.substring(1);
+                        Method method = object.getClass().getMethod(methodName);
+                        int featureValue = (int) method.invoke(object);
+                        // 计算条件概率
+                        posteriorProb *= conditionalProbabilities.get(feature).get(classValue).getOrDefault(featureValue, 0.0);
+                    }
+                }
+                posteriorProbabilities.put(classValue, posteriorProb);
+            }
+
+            // 找到后验概率最大的类别
+            int maxClass = -1;
+            double maxProb = -1;
+            for (Map.Entry<Integer, Double> entry : posteriorProbabilities.entrySet()) {
+                if (entry.getValue() > maxProb) {
+                    maxProb = entry.getValue();
+                    maxClass = entry.getKey();
+                }
+            }
+            return maxClass;
+        } catch (Exception e) {
+            e.printStackTrace();
+            return -1;
+        }
+    }
+
+    public void study() {
+
+        // 计算先验概率
+        List<Integer> classValues = dataTable.getTable().get(dataTable.getKey());
+        Map<Integer, Integer> classCounts = new HashMap<>();
+        for (int value : classValues) {
+            classCounts.put(value, classCounts.getOrDefault(value, 0) + 1);
+        }
+        for (Map.Entry<Integer, Integer> entry : classCounts.entrySet()) {
+            priorProbabilities.put(entry.getKey(), (double) entry.getValue() / dataTable.getLength());
+        }
+
+        // 计算条件概率
+        for (String feature : dataTable.getKeyType()) {
+            if (!feature.equals(dataTable.getKey())) {
+                Map<Integer, Map<Integer, Integer>> featureCounts = new HashMap<>();
+                List<Integer> featureValues = dataTable.getTable().get(feature);
+                for (int i = 0; i < dataTable.getLength(); i++) {
+                    int classValue = classValues.get(i);
+                    int featureValue = featureValues.get(i);
+                    featureCounts.computeIfAbsent(classValue, k -> new HashMap<>()).put(featureValue, featureCounts.get(classValue).getOrDefault(featureValue, 0) + 1);
+                }
+                Map<Integer, Map<Integer, Double>> featureProbabilities = new HashMap<>();
+                for (Map.Entry<Integer, Map<Integer, Integer>> entry : featureCounts.entrySet()) {
+                    int classValue = entry.getKey();
+                    Map<Integer, Integer> counts = entry.getValue();
+                    Map<Integer, Double> probabilities = new HashMap<>();
+                    for (Map.Entry<Integer, Integer> countEntry : counts.entrySet()) {
+                        int featureValue = countEntry.getKey();
+                        int count = countEntry.getValue();
+                        probabilities.put(featureValue, (double) count / classCounts.get(classValue));
+                    }
+                    featureProbabilities.put(classValue, probabilities);
+                }
+                conditionalProbabilities.put(feature, featureProbabilities);
+            }
+        }
+    }
+}
+