def model_predict():
path_config = global_config.PathConfig()
param_config = global_config.ParamConfig()
param_config.set_begin_time('2012-01')
param_config.set_end_time('2019-5')
param_config.set_predict_period(1)
fca = feature_correlation_analysis.FeatureCorrelationAnalysis(
path_config, param_config)
fca.feature_correlation_analysis()
data_coversion.DataCoversion(path_config, param_config).data_coversion()
data_partitioning.DataPartition(path_config,
param_config).data_partitioning()
feature_df = fca.initial_attr_data()
predict_time = datetime.datetime.strptime('2019-6', '%Y-%m')
data_attr = []
with open(path_config.attr_intro, mode='r') as f:
for line in f:
attr, period = line.split()
data_attr.append(
feature_df[attr][predict_time -
dateutil.relativedelta.relativedelta(
months=int(period))])
data_array = np.array(data_attr)
data_array = data_array[np.newaxis, :]
model = BaggingRegression_model.final_Bagging_regression()
data_value = model.predict(data_array)
print(data_value)
def initial_data(data_type=1):
# 用于初始化数据集
path_config = global_config.PathConfig()
param_config = global_config.ParamConfig()
param_config.data_type = data_type # 数据类型
param_config.predict_period = 1 # 预测期
feature_correlation_analysis.FeatureCorrelationAnalysis(
path_config, param_config).feature_correlation_analysis()
data_coversion.DataCoversion(path_config, param_config).data_coversion()
def __init__(self,
path_config=global_config.PathConfig(),
param_config=global_config.ParamConfig()):
self.path_config = path_config
self.param_config = param_config
def __init__(self, path_config=global_config.PathConfig(), param_config=global_config.ParamConfig()):
self.path_config = path_config
self.param_config = param_config
self.MAX_PERIOD = self.param_config.corr_max_period
self.PREDICT_PERIOD = self.param_config.predict_period
"""
Time： 2019-7-112
Description： 一些通用方法
"""
from SteelDemandAnalysis.config import global_config
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import mean_absolute_error
from sklearn.metrics import mean_squared_error
config = global_config.PathConfig()
def load_data(train_data_df, test_data_df):
# 加载训练与验证数据
# 返回：
# train_attr:训练集属性
# train_label:训练集标签
# test_attr:测试集属性
# test_label:测试集标签
train_data = train_data_df.values
train_attr = train_data[:, :-1]
train_label = train_data[:, -1]
test_data = test_data_df.values
test_attr = test_data[:, :-1]
test_label = test_data[:, -1]
return train_attr, train_label, test_attr, test_label
