画像时间衰减上线

eda/smart_rank/dist_update_user_portrait_service.py

@@ -94,11 +94,11 @@ def get_user_service_portrait(cl_id, all_word_tags, all_tag_tag_type, all_3tag_2
             user_df_service["tag2_type"] = user_df_service.apply(lambda x: all_tag_tag_type.get(x["tag2"]), axis=1)
             # 算分及比例
             user_df_service["tag_score"] = user_df_service.apply(
-                lambda x: compute_henqiang(x.days_diff_now)/get_action_tag_count(user_df_service, x.time) if x.score_type == "henqiang" else (
-                    compute_jiaoqiang(x.days_diff_now)/get_action_tag_count(user_df_service, x.time) if x.score_type == "jiaoqiang" else (
-                        compute_ai_scan(x.days_diff_now)/get_action_tag_count(user_df_service, x.time) if x.score_type == "ai_scan" else (
-                        compute_ruoyixiang(x.days_diff_now)/get_action_tag_count(user_df_service, x.time) if x.score_type == "ruoyixiang" else
-                        compute_validate(x.days_diff_now)/get_action_tag_count(user_df_service, x.time)))), axis=1)
+                lambda x: compute_henqiang(x.days_diff_now, exponential=1)/get_action_tag_count(user_df_service, x.time) if x.score_type == "henqiang" else (
+                    compute_jiaoqiang(x.days_diff_now, exponential=1)/get_action_tag_count(user_df_service, x.time) if x.score_type == "jiaoqiang" else (
+                        compute_ai_scan(x.days_diff_now, exponential=1)/get_action_tag_count(user_df_service, x.time) if x.score_type == "ai_scan" else (
+                        compute_ruoyixiang(x.days_diff_now, exponential=1)/get_action_tag_count(user_df_service, x.time) if x.score_type == "ruoyixiang" else
+                        compute_validate(x.days_diff_now, exponential=1)/get_action_tag_count(user_df_service, x.time)))), axis=1)
             tag_score_sum = user_df_service.groupby(by=["tag2", "tag2_type"]).agg(
                 {'tag_score': 'sum', 'cl_id': 'first', 'action': 'first'}).reset_index().sort_values(by=["tag_score"],
                                                                                                      ascending=False)

eda/smart_rank/evaluation_metrics.py

@@ -32,7 +32,7 @@ def setup_logger(logger_name, log_file, level=logging.INFO):
 
 
 def get_user_service_portrait_not_alipay(cl_id, all_word_tags, all_tag_tag_type, pay_time, all_3tag_2tag, version=1,
-                                             exponential=0, normalization_size=7, decay_days=180, size=10):
+                                             exponential=0, normalization_size=7, decay_days=30, size=10):
     """
     :param cl_id:
     :param all_word_tags:
@@ -305,7 +305,7 @@ if __name__ == '__main__':
         parser.add_argument("-e", "--exponential", type=int, dest="exponential", default=0, help="是否采用指数衰减")
         parser.add_argument("-n", "--normalization_size", type=int, dest="normalization_size", default=7,
                             help="天数差归一化的区间")
-        parser.add_argument("-d", "--decay_days", type=int, dest="decay_days", default=180, help="分数衰减的天数")
+        parser.add_argument("-d", "--decay_days", type=int, dest="decay_days", default=30, help="分数衰减的天数")
         parser.add_argument("-a", "--action_type", dest="action_type", nargs='+', help="计算匹配度的行为")
         parser.add_argument("-s", "--save_tidb", type=int, dest="save_tidb", default=1, help="统计结果是否存tidb")
         args = parser.parse_args()

eda/smart_rank/tool.py

@@ -187,7 +187,7 @@ def get_tag2_from_tag3(tag3, all_3tag_2tag, user_log_df_tag2_list):
         print(e)
 
 
-def compute_henqiang(x, decay_days=180, normalization_size=7, exponential=0):
+def compute_henqiang(x, decay_days=30, normalization_size=7, exponential=0):
     if exponential:
         alpha = exponential_decay(x, decay_days, normalization_size)
         score = 15 - 2**alpha * ((15-0.5)/decay_days)
@@ -197,7 +197,7 @@ def compute_henqiang(x, decay_days=180, normalization_size=7, exponential=0):
         return score
     else:
         return 0.5
-def compute_jiaoqiang(x, decay_days=180, normalization_size=7, exponential=0):
+def compute_jiaoqiang(x, decay_days=30, normalization_size=7, exponential=0):
     if exponential:
         alpha = exponential_decay(x, decay_days, normalization_size)
         score = 12 - 2**alpha * ((12-0.5)/decay_days)
@@ -207,7 +207,7 @@ def compute_jiaoqiang(x, decay_days=180, normalization_size=7, exponential=0):
         return score
     else:
         return 0.5
-def compute_ruoyixiang(x, decay_days=180, normalization_size=7, exponential=0):
+def compute_ruoyixiang(x, decay_days=30, normalization_size=7, exponential=0):
     if exponential:
         alpha = exponential_decay(x, decay_days, normalization_size)
         score = 5 - 2**alpha * ((5-0.5)/decay_days)
@@ -217,7 +217,7 @@ def compute_ruoyixiang(x, decay_days=180, normalization_size=7, exponential=0):
         return score
     else:
         return 0.5
-def compute_validate(x, decay_days=180, normalization_size=7, exponential=0):
+def compute_validate(x, decay_days=30, normalization_size=7, exponential=0):
     if exponential:
         alpha = exponential_decay(x, decay_days, normalization_size)
         score = 10 - 2**alpha * ((10-0.5)/decay_days)
@@ -227,7 +227,7 @@ def compute_validate(x, decay_days=180, normalization_size=7, exponential=0):
         return score
     else:
         return 0.5
-def compute_ai_scan(x, decay_days=180, normalization_size=7, exponential=0):
+def compute_ai_scan(x, decay_days=30, normalization_size=7, exponential=0):
     if exponential:
         alpha = exponential_decay(x, decay_days, normalization_size)
         score = 2 - 2**alpha * ((2-0.5)/decay_days)
@@ -247,7 +247,7 @@ def get_action_tag_count(df, action_time):
         print(e)
 
 
-def exponential_decay(days_diff, decay_days=180, normalization_size=7):
+def exponential_decay(days_diff, decay_days=30, normalization_size=7):
     x = np.arange(1, decay_days+1, 1)
     # 天数差归一化到[0, normalization_size]
     a = (normalization_size - 0) * (days_diff - min(x)) / (max(x) - min(x))