l666661的博客

模型融合：1.特征工程多方案2.vote投票机制3.stacking多层融合（一层xgb,lgb,cnn，二层lr）目前在调参阶段出现欠拟合状况，听了前天的鱼佬的讲解，调整了一下方向，结合今天的直播准备先做一下模型改造，然后融合看一下具体的预测情况。...

贝叶斯调参1.数据读取与转换import pandas as pdimport numpy as npfrom sklearn.metrics import f1_score, make_scorerimport osimport seaborn as snsimport matplotlib.pyplot as pltimport warningswarnings.filterwarnings("ignore")def reduce_mem_usage(df):

特征提取：def feature(): # 数据读取 data_train = pd.read_csv('../data/train.csv') test=pd.read_csv('../data/testA.csv') # print(train.head()) # 对心电特征进行行转列处理，同时为每个心电信号加入时间步特征time train_heartbeat_df = data_train["heartbeat_signals"].str.sp