liurui/PriceForecast
Template
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

聚烯烃每周最佳模型逻辑调整

Browse Source
This commit is contained in:
workpc 2025-04-21 17:33:12 +08:00
parent 8074647329
commit 5a8e3ae007
5 changed files with 61 additions and 59 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
config_jingbo_yuedu.py
View File

@ -39,35 +39,35 @@ edbnamelist = [
]
# eta自有数据指标编码
# eta自有数据指标编码 次月
modelsindex = {
'NHITS': 'SELF0000001',
'Informer': 'SELF0000057',
'LSTM': 'SELF0000058',
'iTransformer': 'SELF0000059',
'TSMixer': 'SELF0000060',
'TSMixerx': 'SELF0000061',
'PatchTST': 'SELF0000062',
'RNN': 'SELF0000063',
'GRU': 'SELF0000064',
'TCN': 'SELF0000065',
'BiTCN': 'SELF0000066',
'DilatedRNN': 'SELF0000067',
'MLP': 'SELF0000068',
'DLinear': 'SELF0000069',
'NLinear': 'SELF0000070',
'TFT': 'SELF0000071',
'FEDformer': 'SELF0000072',
'StemGNN': 'SELF0000073',
'MLPMultivariate': 'SELF0000074',
'TiDE': 'SELF0000075',
'DeepNPTS': 'SELF0000076'
'NHITS': 'SELF0000143',
'Informer': 'SELF0000144',
'LSTM': 'SELF0000145',
'iTransformer': 'SELF0000146',
'TSMixer': 'SELF0000147',
'TSMixerx': 'SELF0000148',
'PatchTST': 'SELF0000149',
'RNN': 'SELF0000150',
'GRU': 'SELF0000151',
'TCN': 'SELF0000152',
'BiTCN': 'SELF0000153',
'DilatedRNN': 'SELF0000154',
'MLP': 'SELF0000155',
'DLinear': 'SELF0000156',
'NLinear': 'SELF0000157',
'TFT': 'SELF0000158',
'FEDformer': 'SELF0000159',
'StemGNN': 'SELF0000160',
'MLPMultivariate': 'SELF0000161',
'TiDE': 'SELF0000162',
'DeepNPTS': 'SELF0000163',
}
# eta 上传预测结果的请求体，后面发起请求的时候更改 model datalist 数据
data = {
"IndexCode": "",
"IndexName": "价格预测模型",
"IndexName": "价格预测xx模型-次月",
"Unit": "无",
"Frequency": "日度",
"SourceName": f"价格预测",
@ -305,7 +305,7 @@ is_timefurture = True # 是否使用时间特征
is_fivemodels = False # 是否使用之前保存的最佳的5个模型
is_edbcode = False # 特征使用edbcoding列表中的
is_edbnamelist = False # 自定义特征，对应上面的edbnamelist
is_update_eta = False # 预测结果上传到eta
is_update_eta = True # 预测结果上传到eta
is_update_report = True # 是否上传报告
is_update_warning_data = False # 是否上传预警数据
is_update_predict_value = True # 是否上传预测值到市场信息平台

44
config_jingbo_zhoudu.py
View File

@ -39,29 +39,29 @@ edbnamelist = [
]
# eta自有数据指标编码
# eta自有数据指标编码 次周
modelsindex = {
'NHITS': 'SELF0000001',
'Informer': 'SELF0000057',
'LSTM': 'SELF0000058',
'iTransformer': 'SELF0000059',
'TSMixer': 'SELF0000060',
'TSMixerx': 'SELF0000061',
'PatchTST': 'SELF0000062',
'RNN': 'SELF0000063',
'GRU': 'SELF0000064',
'TCN': 'SELF0000065',
'BiTCN': 'SELF0000066',
'DilatedRNN': 'SELF0000067',
'MLP': 'SELF0000068',
'DLinear': 'SELF0000069',
'NLinear': 'SELF0000070',
'TFT': 'SELF0000071',
'FEDformer': 'SELF0000072',
'StemGNN': 'SELF0000073',
'MLPMultivariate': 'SELF0000074',
'TiDE': 'SELF0000075',
'DeepNPTS': 'SELF0000076'
'NHITS': 'SELF0000099',
'Informer': 'SELF0000100',
'LSTM': 'SELF0000101',
'iTransformer': 'SELF0000102',
'TSMixer': 'SELF0000103',
'TSMixerx': 'SELF0000104',
'PatchTST': 'SELF0000105',
'RNN': 'SELF0000106',
'GRU': 'SELF0000107',
'TCN': 'SELF0000108',
'BiTCN': 'SELF0000109',
'DilatedRNN': 'SELF0000110',
'MLP': 'SELF0000111',
'DLinear': 'SELF0000112',
'NLinear': 'SELF0000113',
'TFT': 'SELF0000114',
'FEDformer': 'SELF0000115',
'StemGNN': 'SELF0000116',
'MLPMultivariate': 'SELF0000117',
'TiDE': 'SELF0000118',
'DeepNPTS': 'SELF0000119',
}
# eta 上传预测结果的请求体，后面发起请求的时候更改 model datalist 数据

3
config_juxiting.py
View File

@ -219,7 +219,6 @@ push_data_value_list_data = {
}
# 八大维度数据项编码
bdwd_items = {
'ciri': 'jxtppbdwdcr',
@ -243,7 +242,7 @@ table_name = 'v_tbl_crude_oil_warning'
# 开关
is_train = True # 是否训练
is_debug = False # 是否调试
is_eta = True # 是否使用eta接口
is_eta = False # 是否使用eta接口
is_market = False # 是否通过市场信息平台获取特征 ,在is_eta 为true 的情况下生效
is_timefurture = True # 是否使用时间特征
is_fivemodels = False # 是否使用之前保存的最佳的5个模型

11
main_juxiting.py
View File

@ -302,12 +302,13 @@ def predict_main():
# 计算最近60天预测残差最低的模型名称
model_results = sqlitedb.select_data(
'trueandpredict', order_by="ds DESC", limit="60")
# 删除created_dt,y列空行
model_results = model_results.dropna(
subset=['created_dt', 'y'])
# 删除空值率为90%以上的列
if len(model_results) > 10:
model_results = model_results.dropna(
thresh=len(model_results)*0.1, axis=1)
# 删除空行
model_results = model_results.dropna()
model_results = model_results.dropna(
thresh=len(model_results)*0.1, axis=1)
modelnames = model_results.columns.to_list()[2:-1]
for col in model_results[modelnames].select_dtypes(include=['object']).columns:
model_results[col] = model_results[col].astype(np.float32)

14
models/nerulforcastmodels.py
View File

@ -250,7 +250,8 @@ def ex_Model(df, horizon, input_size, train_steps, val_check_steps, early_stop_p
list.append({'Date': date, 'Value': value})
data['DataList'] = list
data['IndexCode'] = modelsindex[m]
data['IndexName'] = f'价格预测{m}模型'
# data['IndexName'] = f'价格预测{m}模型'
data['IndexName'] = data['IndexName'].replece('xx', m)
data['Remark'] = m
etadata.push_data(data)
@ -463,7 +464,7 @@ def ex_Model_Juxiting(df, horizon, input_size, train_steps, val_check_steps, ear
if config.is_update_eta:
df_predict['ds'] = pd.to_datetime(df_predict['ds'])
dates = df_predict['ds'].dt.strftime('%Y-%m-%d')
for m in modelsindex.keys():
try:
list = []
@ -1247,7 +1248,6 @@ def model_losss(sqlitedb, end_time):
_get_accuracy_rate(df, create_dates, ds_dates)
except Exception as e:
config.logger.info(f'准确率计算错误{e}')
def _add_abs_error_rate():
# 计算每个预测值与真实值之间的偏差率
@ -1822,7 +1822,8 @@ def model_losss_juxitingbak(sqlitedb, end_time):
plt.text(i, j, str(j), ha='center', va='bottom')
# 当前日期画竖虚线
plt.axvline(x=df['ds'].iloc[-config.horizon], color='r', linestyle='--')
plt.axvline(x=df['ds'].iloc[-config.horizon],
color='r', linestyle='--')
plt.legend()
plt.xlabel('日期')
# 自动设置横轴日期显示
@ -2270,7 +2271,7 @@ def brent_export_pdf(num_indicators=475, num_models=21, num_dayindicator=202, in
config.dataset, 'model_evaluation.csv'), encoding='utf-8')
# 按评估指标排序，取前五
fivemodels_list = eval_df['模型(Model)'].values # 列表形式，后面当作列名索引使用
eval_models_list = fivemodels_list.tolist()
eval_models_list = fivemodels_list.tolist()
# 去掉FEDformer
if 'FEDformer' in eval_models_list:
eval_models_list.remove('FEDformer')
@ -2356,7 +2357,8 @@ def brent_export_pdf(num_indicators=475, num_models=21, num_dayindicator=202, in
content.append(Graphs.draw_little_title('模型选择：'))
content.append(Graphs.draw_text(
f'本次预测使用了一个专门收集时间序列的NeuralForecast库中的{num_models}个模型：'))
content.append(Graphs.draw_text(f'使用{config.input_size}个数据预测未来{inputsize}个数据。'))
content.append(Graphs.draw_text(
f'使用{config.input_size}个数据预测未来{inputsize}个数据。'))
content.append(Graphs.draw_little_title('指标情况：'))
with open(os.path.join(config.dataset, '特征频度统计.txt'), encoding='utf-8') as f:
for line in f.readlines():