| 提交 | 用户 | 时间 | ||
| 449017 | 1 | package com.iailab.module.model.mpk.common.utils; |
| D | 2 | |
| 3 | import java.util.HashMap; | |
| 4 | ||
| 5 | ||
| 6 | /** | |
| 7 | * @Auther: Forrest | |
| 8 | * @Date: 2020/6/8 14:05 | |
| 9 | * @Description: | |
| 10 | */ | |
| 11 | public class AlgsUtils { | |
| 12 | private HashMap<String, Object> model = new HashMap<String, Object>(); | |
| 13 | ||
| 14 | public HashMap<String, Object> createPredictHashmap(HashMap<String, Object> models) { | |
| 15 | if ((models.containsKey("iail/mdk/model"))) { | |
| 16 | if (((String) ((HashMap) models.get("iail/mdk/model")).get("param1")).isEmpty()) { | |
| 17 | String aaa = "error"; | |
| 18 | model.put("param1", aaa); | |
| 19 | } else { | |
| 20 | String model_train = (String) ((HashMap) models.get("iail/mdk/model")).get("param1"); | |
| 21 | model.put("param1", model_train); | |
| 22 | } | |
| 23 | } else { | |
| 24 | model = models; | |
| 25 | } | |
| 26 | return model; | |
| 27 | } | |
| 28 | ||
| 29 | public HashMap<String, Object> createPredictHashmapplus(HashMap<String, Object> models) { | |
| 30 | if (models != null && models.containsKey("models")) { | |
| 31 | if (((String) ((HashMap) models.get("models")).get("paramFile")).isEmpty()) { | |
| 32 | String aaa = "error"; | |
| 33 | model.put("param1", aaa); | |
| 34 | } else { | |
| 35 | String model_train = (String) ((HashMap) models.get("models")).get("paramFile"); | |
| 36 | model.put("paramFile", model_train); | |
| 37 | if (((HashMap) models.get("models")).containsKey("dim")) { | |
| 38 | Object dim = ((HashMap) models.get("models")).get("dim"); | |
| 39 | model.put("dim", dim); | |
| 40 | } | |
| 41 | } | |
| 42 | } else { | |
| 43 | model = models; | |
| 44 | } | |
| 45 | return model; | |
| 46 | } | |
| 47 | ||
| 48 | private HashMap<String, Object> eval_pre = new HashMap<String, Object>(); | |
| 49 | ||
| 50 | /** | |
| 51 | * 对返回码进行转换 | |
| 52 | * | |
| 53 | * @param models | |
| 54 | * @return | |
| 55 | */ | |
| 56 | public int reverseResultCode(HashMap<String, Object> models) { | |
| 57 | if ((models.containsKey("result_code"))) { | |
| 58 | return Integer.parseInt((String) models.get("result_code")); | |
| 59 | } | |
| 60 | return -2; | |
| 61 | } | |
| 62 | ||
| 63 | /** | |
| 64 | * 对评价指标进行转换 | |
| 65 | * | |
| 66 | * @param models | |
| 67 | * @return | |
| 68 | */ | |
| 69 | public HashMap<String, Object> reverseEval(HashMap<String, Object> models) { | |
| 70 | if ((models.containsKey("eval"))) { | |
| 71 | if (((HashMap) models.get("eval")).containsKey("MAE")) { | |
| 72 | double MAE = Double.parseDouble((String) ((HashMap) models.get("eval")).get("MAE")); | |
| 73 | eval_pre.put("MAE", MAE); | |
| 74 | } | |
| 75 | if (((HashMap) models.get("eval")).containsKey("MAPE")) { | |
| 76 | double MAPE = Double.parseDouble((String) ((HashMap) models.get("eval")).get("MAPE")); | |
| 77 | eval_pre.put("MAPE", MAPE); | |
| 78 | } | |
| 79 | if (((HashMap) models.get("eval")).containsKey("RMSE")) { | |
| 80 | double MAE = Double.parseDouble((String) ((HashMap) models.get("eval")).get("RMSE")); | |
| 81 | eval_pre.put("RMSE", MAE); | |
| 82 | } | |
| 83 | } | |
| 84 | ||
| 85 | return eval_pre; | |
| 86 | } | |
| 87 | ||
| 88 | /** | |
| 89 | * 对models里面的参数进行转换 | |
| 90 | */ | |
| 91 | private HashMap<String, Object> train_result_models = new HashMap<String, Object>(); | |
| 92 | ||
| 93 | public HashMap<String, Object> reverseModels(HashMap<String, Object> train_result) { | |
| 94 | if (train_result.containsKey("models")) { | |
| 95 | train_result_models = (HashMap) train_result.get("models"); | |
| 96 | if (((HashMap) train_result.get("models")).containsKey("dim")) { | |
| 97 | double dim = Double.parseDouble((String) ((HashMap) train_result.get("models")).get("dim")); | |
| 98 | train_result_models.put("dim", dim); | |
| 99 | } | |
| 100 | train_result.put("models", train_result_models); | |
| 101 | } | |
| 102 | return train_result; | |
| 103 | } | |
| 104 | ||
| 105 | ||
| 106 | /** | |
| 107 | * 获取二维数组行列数 | |
| 108 | * | |
| 109 | * @param arr | |
| 110 | * @return | |
| 111 | */ | |
| 112 | public int[] getColAndRow(double[][] arr) { | |
| 113 | int row = arr.length; | |
| 114 | int col = arr[0].length; | |
| 115 | int[] result = new int[2]; | |
| 116 | result[0] = row; | |
| 117 | result[1] = col; | |
| 118 | return result; | |
| 119 | } | |
| 120 | ||
| 121 | /** | |
| 122 | * 两个二维数组进行合并 | |
| 123 | * | |
| 124 | * @param data | |
| 125 | * @param refs | |
| 126 | * @return | |
| 127 | */ | |
| 128 | public double[][] getMathergeArr(double[][] data, double[][] refs) { | |
| 129 | ||
| 130 | int[] dataRowAndCol = getColAndRow(data); | |
| 131 | int rowData = dataRowAndCol[0]; | |
| 132 | int colData = dataRowAndCol[1]; | |
| 133 | ||
| 134 | int[] refsRowAndCol = getColAndRow(refs); | |
| 135 | int rowrefs = refsRowAndCol[0]; | |
| 136 | int colrefs = refsRowAndCol[1]; | |
| 137 | ||
| 138 | double[][] newData = new double[rowData + rowrefs][colData]; | |
| 139 | for (int i = 0; i < rowData; i++) { | |
| 140 | for (int j = 0; j < colData; j++) { | |
| 141 | newData[i][j] = data[i][j]; | |
| 142 | } | |
| 143 | } | |
| 144 | ||
| 145 | for (int i = 0; i < rowrefs; i++) { | |
| 146 | for (int j = 0; j < colrefs; j++) { | |
| 147 | newData[i + rowData][j] = refs[i][j]; | |
| 148 | } | |
| 149 | } | |
| 150 | return newData; | |
| 151 | } | |
| 152 | ||
| 153 | /** | |
| 154 | * 对训练方法进行处理,实现评价指标的转换 | |
| 155 | */ | |
| 156 | public HashMap<String, Object> trainUtil(HashMap<String, Object> train_result, HashMap<String, Object> eval, String time) { | |
| 157 | if (train_result.containsKey("eval")) { | |
| 158 | eval = (HashMap<String, Object>) train_result.get("eval"); | |
| 159 | eval.put("time", time); | |
| 160 | train_result.put("eval", eval); | |
| 161 | } | |
| 162 | train_result.put("result_code", reverseResultCode(train_result)); | |
| 163 | return train_result; | |
| 164 | } | |
| 165 | ||
| 166 | /** | |
| 167 | * 对预测方法进行处理 | |
| 168 | */ | |
| 169 | // public HashMap<String,Object> predictUtil(){ | |
| 170 | // | |
| 171 | // } | |
| 172 | } | |
