本文共 2402 字,大约阅读时间需要 8 分钟。
下一篇:
预测任务:用户是否会下载APP,当其点击广告以后
数据集:ks-projects-201801.csv'deadline','launched'
,parse_dates
解析为时间ks = pd.read_csv('ks-projects-201801.csv',parse_dates=['deadline','launched'])预测Kickstarter项目是否会成功。
state
作为结果label
可以使用类别category
,货币currency
,资金目标funding goal
,国家country
以及启动时间launched
等特征 pd.unique(ks.state)
有6种数值
array(['failed', 'canceled', 'successful', 'live', 'undefined', 'suspended'], dtype=object)
每种多少个?按state
分组,每组中ID
行数有多少
ks.groupby('state')['ID'].count()
statecanceled 38779failed 197719live 2799successful 133956suspended 1846undefined 3562Name: ID, dtype: int64
live
丢弃,successful
的标记为1,其余的为0ks = ks.query('state != "live"') # live行不要ks = ks.assign(outcome=(ks['state']=='successful').astype(int))# label 转成1,0,int型
launched
时间拆分成,年月日小时,作为新的特征ks = ks.assign(hour=ks.launched.dt.hour, day=ks.launched.dt.day, month=ks.launched.dt.month, year=ks.launched.dt.year)ks.head()
category, currency, country
为数字from sklearn.preprocessing import LabelEncodercat_features = ['category','currency','country']encoder = LabelEncoder()encoded = ks[cat_features].apply(encoder.fit_transform)encoded.head(10)
X = ks[['goal', 'hour', 'day', 'month', 'year', 'outcome']].join(encoded)X.head()
sklearn.model_selection.StratifiedShuffleSplit
valid_ratio = 0.1valid_size = int(len(X)*valid_ratio)train = X[ : -2*valid_size]valid = X[-2*valid_size : -valid_size]test = X[-valid_size : ]
需要关注下,label 在每个数据集中的占比是否接近
for each in [train, valid, test]: print("Outcome fraction = {:.4f}".format(each.outcome.mean()))
Outcome fraction = 0.3570Outcome fraction = 0.3539Outcome fraction = 0.3542
feature_cols = train.columns.drop('outcome')dtrain = lgb.Dataset(train[feature_cols], label=train['outcome'])dvalid = lgb.Dataset(valid[feature_cols], label=valid['outcome'])param = { 'num_leaves': 64, 'objective': 'binary'}param['metric'] = 'auc'num_round = 1000bst = lgb.train(param, dtrain, num_round, valid_sets=[dvalid], early_stopping_rounds=10, verbose_eval=False)
from sklearn import metricsypred = bst.predict(test[feature_cols])score = metrics.roc_auc_score(test['outcome'], ypred)print(f"Test AUC score: {score}")
下一篇:
转载地址:http://nuhtf.baihongyu.com/