补充标准化

src/main/java/io/lroyia/model/knn/KNNModel.java

@@ -56,9 +56,10 @@ public class KNNModel<ResultT> implements ClassifyModel<ResultT> {
         double[][] nTrainData = new double[trainData.length][];
         this.maxArray = new double[trainData.length];
         for (int i = 0; i < trainData.length; i++) {
-            DataUtil.ToOneResult toOneResult = DataUtil.columnToOne(trainData[i]);
+            DataUtil.MaxMin maxMin = DataUtil.getMaxMin(trainData[i]);
-            this.maxArray[i] = toOneResult.getMax();
+            this.maxArray[i] = maxMin.getMax();
-            this.minArray[i] = toOneResult.getMin();
+            this.minArray[i] = maxMin.getMin();
+            DataUtil.ToOneResult toOneResult = DataUtil.columnToOne(DataUtil.standardData(trainData[i]));
             nTrainData[i] = toOneResult.getResult();
 
         }

src/main/java/io/lroyia/util/DataUtil.java

@@ -86,6 +86,7 @@ public abstract class DataUtil {
         return result;
     }
 
+
     /**
      * 列归一
      *
@@ -95,12 +96,9 @@ public abstract class DataUtil {
      * @since 2023年10月24日 10:27:05
      */
     public static ToOneResult columnToOne(double[] arr) {
-        double max = Double.MIN_VALUE;
+        MaxMin maxMin = getMaxMin(arr);
-        double min = Double.MAX_VALUE;
+        double max = maxMin.max;
-        for (Double each : arr) {
+        double min = maxMin.min;
-            max = Math.max(each, max);
-            min = Math.min(each, min);
-        }
         double divisor = max - min;
         double[] result = new double[arr.length];
         for (int i = 0; i < arr.length; i++) {
@@ -110,6 +108,50 @@ public abstract class DataUtil {
         return new ToOneResult(max, min, result);
     }
 
+    /**
+     * 获取最大最小值
+     *
+     * @param arr 数组
+     * @return 最大最小值
+     * @author lroyia
+     * @since 2023年10月29日 16:09:38
+     */
+    public static MaxMin getMaxMin(double[] arr) {
+        double max = Double.MIN_VALUE;
+        double min = Double.MAX_VALUE;
+        for (Double each : arr) {
+            max = Math.max(each, max);
+            min = Math.min(each, min);
+        }
+        return new MaxMin(max, min);
+    }
+
+    /**
+     * 列归一
+     *
+     * @param arr 列数据
+     * @return 归一
+     * @author lroyia
+     * @since 2023年10月24日 10:27:05
+     */
+    public static double[] standardData(double[] arr) {
+        double[] result = new double[arr.length];
+        double sum = 0;
+        for (double each : arr) {
+            sum += each;
+        }
+        double avg = sum / arr.length;
+        double variance = 0;
+        for (double each : arr) {
+            variance += Math.pow(each - avg, 2);
+        }
+        variance = Math.sqrt(variance);
+        for (int i = 0; i < arr.length; i++) {
+            result[i] = (arr[i] - avg) / variance;
+        }
+        return result;
+    }
+
     /**
      * 计算距离
      *
@@ -127,12 +169,9 @@ public abstract class DataUtil {
         return Math.sqrt(stdDev);
     }
 
-    /**
-     * 归一结果
-     */
     @Getter
     @AllArgsConstructor
-    public static class ToOneResult {
+    public static class MaxMin {
         /**
          * 最大值
          */
@@ -142,10 +181,22 @@ public abstract class DataUtil {
          * 最小值
          */
         private double min;
+    }
+
+    /**
+     * 归一结果
+     */
+    @Getter
+    public static class ToOneResult extends MaxMin {
 
         /**
          * 归一结果
          */
-        private double[] result;
+        private final double[] result;
+
+        public ToOneResult(double max, double min, double[] result) {
+            super(max, min);
+            this.result = result;
+        }
     }
 }