Sfoglia il codice sorgente

feat:修改分析脚本

zhaohaipeng 2 mesi fa
parent
commit
9fd2cd037d

+ 3 - 0
src/main/java/examples/extractor/ExtractorUtils.java

@@ -176,6 +176,9 @@ public class ExtractorUtils {
     }
 
     public static double reciprocal(double num) {
+        if (num == 0) {
+            return 0;
+        }
         return 1.0 / (num + 1);
     }
 

+ 3 - 3
src/main/java/examples/extractor/v20250218/ExtractFeature20250218.java

@@ -67,7 +67,7 @@ public class ExtractFeature20250218 {
                 featureMap.put("b12_" + index + "_" + time, value);
             }
             double rovn = Double.parseDouble(b12Feature.getOrDefault("rovn_" + time, "0").toString());
-            double returnNUv = Double.parseDouble(b12Feature.getOrDefault("return_n_uv_"+time, "0").toString());
+            double returnNUv = Double.parseDouble(b12Feature.getOrDefault("return_n_uv_" + time, "0").toString());
             featureMap.put("b12_rovn*log(r)_" + time, rovn * RankExtractorFeature_20240530.calLog(returnNUv));
         }
     }
@@ -99,7 +99,7 @@ public class ExtractFeature20250218 {
         featureMap.put("hour", ExtractorUtils.getHourByTimestamp(ts));
         featureMap.put("day_of_week", ExtractorUtils.getDayOfWeekByTimestamp(ts));
 
-        long createTs = Long.parseLong(videoFeature.getOrDefault("gmt_create_timestamp", "0").toString());
+        long createTs = Long.parseLong(videoFeature.getOrDefault("gmt_create_timestamp", "0").toString()) / 1000;
         featureMap.put("create_ts_diff", ExtractorUtils.getDaysBetween(createTs, ts));
 
         String date = LocalDateTime.ofInstant(Instant.ofEpochSecond(ts), ZoneId.systemDefault()).format(DateTimeFormatter.ofPattern("yyyy-MM-dd"));
@@ -124,7 +124,7 @@ public class ExtractFeature20250218 {
                 featureMap.put("c1_" + index + "_" + time, value);
             }
             double rovn = Double.parseDouble(c1Feature.getOrDefault("rovn_" + time, "0").toString());
-            double returnNUv = Double.parseDouble(c1Feature.getOrDefault("return_1_uv_"+time, "0").toString());
+            double returnNUv = Double.parseDouble(c1Feature.getOrDefault("return_1_uv_" + time, "0").toString());
             featureMap.put("c1_rovn*log(r)_" + time, rovn * RankExtractorFeature_20240530.calLog(returnNUv));
         }
     }

+ 94 - 94
src/main/resources/feature_name_20250218.txt

@@ -40,13 +40,13 @@ b1_ros_one_24h
 b1_ros_one_3h
 b1_ros_one_6h
 b1_ros_one_72h
-b1_rovn
-b1_rovn
-b1_rovn
-b1_rovn
-b1_rovn
-b1_rovn
-b1_rovn
+b1_rovn*log(r)_12h
+b1_rovn*log(r)_168h
+b1_rovn*log(r)_1h
+b1_rovn*log(r)_24h
+b1_rovn*log(r)_3h
+b1_rovn*log(r)_6h
+b1_rovn*log(r)_72h
 b1_rovn_12h
 b1_rovn_168h
 b1_rovn_1h
@@ -124,13 +124,13 @@ b2_ros_one_24h
 b2_ros_one_3h
 b2_ros_one_6h
 b2_ros_one_72h
-b2_rovn
-b2_rovn
-b2_rovn
-b2_rovn
-b2_rovn
-b2_rovn
-b2_rovn
+b2_rovn*log(r)_12h
+b2_rovn*log(r)_168h
+b2_rovn*log(r)_1h
+b2_rovn*log(r)_24h
+b2_rovn*log(r)_3h
+b2_rovn*log(r)_6h
+b2_rovn*log(r)_72h
 b2_rovn_12h
 b2_rovn_168h
 b2_rovn_1h
@@ -208,13 +208,13 @@ b3_ros_one_24h
 b3_ros_one_3h
 b3_ros_one_6h
 b3_ros_one_72h
-b3_rovn
-b3_rovn
-b3_rovn
-b3_rovn
-b3_rovn
-b3_rovn
-b3_rovn
+b3_rovn*log(r)_12h
+b3_rovn*log(r)_168h
+b3_rovn*log(r)_1h
+b3_rovn*log(r)_24h
+b3_rovn*log(r)_3h
+b3_rovn*log(r)_6h
+b3_rovn*log(r)_72h
 b3_rovn_12h
 b3_rovn_168h
 b3_rovn_1h
@@ -292,13 +292,13 @@ b4_ros_one_24h
 b4_ros_one_3h
 b4_ros_one_6h
 b4_ros_one_72h
-b4_rovn
-b4_rovn
-b4_rovn
-b4_rovn
-b4_rovn
-b4_rovn
-b4_rovn
+b4_rovn*log(r)_12h
+b4_rovn*log(r)_168h
+b4_rovn*log(r)_1h
+b4_rovn*log(r)_24h
+b4_rovn*log(r)_3h
+b4_rovn*log(r)_6h
+b4_rovn*log(r)_72h
 b4_rovn_12h
 b4_rovn_168h
 b4_rovn_1h
@@ -376,13 +376,13 @@ b5_ros_one_24h
 b5_ros_one_3h
 b5_ros_one_6h
 b5_ros_one_72h
-b5_rovn
-b5_rovn
-b5_rovn
-b5_rovn
-b5_rovn
-b5_rovn
-b5_rovn
+b5_rovn*log(r)_12h
+b5_rovn*log(r)_168h
+b5_rovn*log(r)_1h
+b5_rovn*log(r)_24h
+b5_rovn*log(r)_3h
+b5_rovn*log(r)_6h
+b5_rovn*log(r)_72h
 b5_rovn_12h
 b5_rovn_168h
 b5_rovn_1h
@@ -460,13 +460,13 @@ b6_ros_one_24h
 b6_ros_one_3h
 b6_ros_one_6h
 b6_ros_one_72h
-b6_rovn
-b6_rovn
-b6_rovn
-b6_rovn
-b6_rovn
-b6_rovn
-b6_rovn
+b6_rovn*log(r)_12h
+b6_rovn*log(r)_168h
+b6_rovn*log(r)_1h
+b6_rovn*log(r)_24h
+b6_rovn*log(r)_3h
+b6_rovn*log(r)_6h
+b6_rovn*log(r)_72h
 b6_rovn_12h
 b6_rovn_168h
 b6_rovn_1h
@@ -544,13 +544,13 @@ b7_ros_one_24h
 b7_ros_one_3h
 b7_ros_one_6h
 b7_ros_one_72h
-b7_rovn
-b7_rovn
-b7_rovn
-b7_rovn
-b7_rovn
-b7_rovn
-b7_rovn
+b7_rovn*log(r)_12h
+b7_rovn*log(r)_168h
+b7_rovn*log(r)_1h
+b7_rovn*log(r)_24h
+b7_rovn*log(r)_3h
+b7_rovn*log(r)_6h
+b7_rovn*log(r)_72h
 b7_rovn_12h
 b7_rovn_168h
 b7_rovn_1h
@@ -628,13 +628,13 @@ b8_ros_one_24h
 b8_ros_one_3h
 b8_ros_one_6h
 b8_ros_one_72h
-b8_rovn
-b8_rovn
-b8_rovn
-b8_rovn
-b8_rovn
-b8_rovn
-b8_rovn
+b8_rovn*log(r)_12h
+b8_rovn*log(r)_168h
+b8_rovn*log(r)_1h
+b8_rovn*log(r)_24h
+b8_rovn*log(r)_3h
+b8_rovn*log(r)_6h
+b8_rovn*log(r)_72h
 b8_rovn_12h
 b8_rovn_168h
 b8_rovn_1h
@@ -712,13 +712,13 @@ b9_ros_one_24h
 b9_ros_one_3h
 b9_ros_one_6h
 b9_ros_one_72h
-b9_rovn
-b9_rovn
-b9_rovn
-b9_rovn
-b9_rovn
-b9_rovn
-b9_rovn
+b9_rovn*log(r)_12h
+b9_rovn*log(r)_168h
+b9_rovn*log(r)_1h
+b9_rovn*log(r)_24h
+b9_rovn*log(r)_3h
+b9_rovn*log(r)_6h
+b9_rovn*log(r)_72h
 b9_rovn_12h
 b9_rovn_168h
 b9_rovn_1h
@@ -796,13 +796,13 @@ b10_ros_one_24h
 b10_ros_one_3h
 b10_ros_one_6h
 b10_ros_one_72h
-b10_rovn
-b10_rovn
-b10_rovn
-b10_rovn
-b10_rovn
-b10_rovn
-b10_rovn
+b10_rovn*log(r)_12h
+b10_rovn*log(r)_168h
+b10_rovn*log(r)_1h
+b10_rovn*log(r)_24h
+b10_rovn*log(r)_3h
+b10_rovn*log(r)_6h
+b10_rovn*log(r)_72h
 b10_rovn_12h
 b10_rovn_168h
 b10_rovn_1h
@@ -880,13 +880,13 @@ b11_ros_one_24h
 b11_ros_one_3h
 b11_ros_one_6h
 b11_ros_one_72h
-b11_rovn
-b11_rovn
-b11_rovn
-b11_rovn
-b11_rovn
-b11_rovn
-b11_rovn
+b11_rovn*log(r)_12h
+b11_rovn*log(r)_168h
+b11_rovn*log(r)_1h
+b11_rovn*log(r)_24h
+b11_rovn*log(r)_3h
+b11_rovn*log(r)_6h
+b11_rovn*log(r)_72h
 b11_rovn_12h
 b11_rovn_168h
 b11_rovn_1h
@@ -946,10 +946,10 @@ b12_ros_one_14d
 b12_ros_one_30d
 b12_ros_one_60d
 b12_ros_one_7d
-b12_rovn
-b12_rovn
-b12_rovn
-b12_rovn
+b12_rovn*log(r)_14d
+b12_rovn*log(r)_30d
+b12_rovn*log(r)_60d
+b12_rovn*log(r)_7d
 b12_rovn_14d
 b12_rovn_30d
 b12_rovn_60d
@@ -1012,13 +1012,13 @@ b13_ros_one_24h
 b13_ros_one_3h
 b13_ros_one_6h
 b13_ros_one_72h
-b13_rovn
-b13_rovn
-b13_rovn
-b13_rovn
-b13_rovn
-b13_rovn
-b13_rovn
+b13_rovn*log(r)_12h
+b13_rovn*log(r)_168h
+b13_rovn*log(r)_1h
+b13_rovn*log(r)_24h
+b13_rovn*log(r)_3h
+b13_rovn*log(r)_6h
+b13_rovn*log(r)_72h
 b13_rovn_12h
 b13_rovn_168h
 b13_rovn_1h
@@ -1054,7 +1054,6 @@ b13_str_plus_24h
 b13_str_plus_3h
 b13_str_plus_6h
 b13_str_plus_72h
-bit_rate
 c1_click_12h
 c1_click_168h
 c1_click_24h
@@ -1083,10 +1082,10 @@ c1_ros_one_12h
 c1_ros_one_168h
 c1_ros_one_24h
 c1_ros_one_72h
-c1_rovn
-c1_rovn
-c1_rovn
-c1_rovn
+c1_rovn*log(r)_12h
+c1_rovn*log(r)_168h
+c1_rovn*log(r)_24h
+c1_rovn*log(r)_72h
 c1_rovn_12h
 c1_rovn_168h
 c1_rovn_24h
@@ -1217,6 +1216,7 @@ d2_score
 d3_exp
 d3_return_n
 d3_rovn
+bit_rate
 day_of_week
 festive_sim
 head_title_festive_sim
@@ -1234,4 +1234,4 @@ today_is_fes
 total_time
 video_fes_eq
 width
-width
+width/height