From c500e497d77abe6aefc52762035360c4017a8d4a Mon Sep 17 00:00:00 2001 From: "shenjian@fangjia.com" Date: Thu, 12 Jan 2023 17:31:38 +0800 Subject: [PATCH] =?UTF-8?q?=E6=A5=BC=E7=9B=98=E8=AF=84=E7=BA=A7=E6=94=B9?= =?UTF-8?q?=E4=B8=BA1-5?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../business/community/price_rating.py | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) diff --git a/dags/business/business/community/price_rating.py b/dags/business/business/community/price_rating.py index d6669c6..2860934 100644 --- a/dags/business/business/community/price_rating.py +++ b/dags/business/business/community/price_rating.py @@ -16,12 +16,13 @@ log = logging.getLogger(__name__) es206_client = ElasticSearchClient(host='192.168.0.206') index_name = "community_summary" + def gmm_train(total_rate, consumption_df, source, target): # 高斯混合模型 - gmm = GaussianMixture(n_components=total_rate).fit(consumption_df[[source]]) - consumption_df['cluster'] = gmm.predict(consumption_df[[source]]) - consumption_df['price_mean_cluster'] = consumption_df.groupby('cluster')[source].transform('mean') - consumption_df[target] = consumption_df['price_mean_cluster'].rank(method='dense').astype(int) + gmm = GaussianMixture(n_components=total_rate, covariance_type='tied', max_iter=200, n_init=10).fit(consumption_df[[source]]) + consumption_df.loc[:, 'cluster'] = gmm.predict(consumption_df[[source]]) + consumption_df.loc[:, 'price_mean_cluster'] = consumption_df.groupby('cluster')[source].transform('mean') + consumption_df.loc[:, target] = consumption_df['price_mean_cluster'].rank(method='dense', ascending=False).astype(int) consumption_df = consumption_df.drop(columns=['cluster', 'price_mean_cluster'], axis=1) return consumption_df @@ -46,12 +47,12 @@ def calc_price_rating(city=None): for city in cities: df = consumption_aim_df[consumption_aim_df['city'] == city].copy() - if len(df) < 10: + if len(df) < 5: continue - df = gmm_train(10, df, source, target) - if len(df) >= 100: - df = gmm_train(100, df, source, score_target) - df['price_score'] /= 10 + df = gmm_train(5, df, source, target) + if len(df) >= 5: + df = gmm_train(5, df, source, score_target) + df['price_score'] /= 5 # 批量更新 for index, item in df.iterrows():