我需要为网络节点的接下来N天建立回归模型。
我的数据集包含以下功能：
YYYYMMDD信息；星期几（影响流量）；客户依赖该节点;是否是假期;和另外两个也会影响流量的参数。
标签：MBPS，是当天最大流量（以mbps指标为单位）。
我尝试了LinearRegression（），RandomForest（），SVR（），AdaBoostRegressor（）以及许多其他方法，但我什至无法在这些模型上获得正r2评分或可接受的MSE或MAE。
能否请您帮我在此数据集上建立基础模型，以便通过使用超参数调整等来做得更好。
我的目标是找到一个匹配的模型，并提前N天生成预测。
谢谢
样本数据集：
"YEAR_I","DAYOW_I","CST","IS_HDAY","MATCH_COUNT","PRIORITY","MBPS"
"20190620","5",167,0,0,0,826279
"20190624","2",169,0,0,0,775120
"20190625","3",168,0,0,0,753918
"20190626","4",168,0,0,0,807338
"20190627","5",168,0,0,0,739691
"20190628","6",168,0,0,0,844546
"20190629","7",162,0,0,0,691837
"20190630","1",162,0,0,0,754769
"20190701","2",162,0,0,0,903428
"20190702","3",161,0,0,0,856882
"20190703","4",158,0,0,0,767688
"20190704","5",158,0,0,0,779772
"20190712","6",150,0,0,0,710226
"20190713","7",156,1,0,0,764292
"20190714","1",153,1,0,0,848846
"20190715","2",154,2,0,0,904547
"20190716","3",154,0,0,0,750170
"20190717","4",153,0,0,0,734258
"20190718","5",154,0,0,0,845613
"20190719","6",157,0,0,0,779454
"20190720","7",156,0,0,0,728447
"20190721","1",153,0,0,0,684583
"20190723","3",151,0,0,0,849301
"20190724","4",151,0,0,0,765968
"20190725","5",151,0,0,0,739072
"20190726","6",150,0,0,0,755601
"20190727","7",155,0,0,0,772812
"20190728","1",155,0,0,0,731838
"20190729","2",155,0,0,0,834939
"20190730","3",153,0,0,0,702836
"20190731","4",152,0,0,0,726787
"20190801","5",152,0,0,0,724125
"20190802","6",153,0,0,0,764430
"20190803","7",148,0,0,0,831460
"20190804","1",148,0,0,0,658365
"20190805","2",143,0,0,0,755955
"20190806","3",142,0,0,0,711848
"20190807","4",142,0,0,0,624254
"20190808","5",143,0,0,0,754606
"20190809","6",141,0,0,0,637655
"20190810","7",134,1,0,0,658406
"20190811","1",130,3,0,0,613327
"20190812","2",124,3,0,0,592067
"20190813","3",124,3,0,0,614577
"20190814","4",125,3,0,0,662439
"20190815","5",130,0,0,0,717342
"20190816","6",137,0,1,2,826435
"20190817","7",141,0,3,2,972885
"20190818","1",140,0,4,2,722316
"20190819","2",142,0,1,2,797723
"20190820","3",147,0,0,0,957704
"20190821","4",146,0,0,0,759117
"20190822","5",144,0,0,0,737167
"20190823","6",149,0,1,2,776062
"20190824","7",148,0,3,1,744867
"20190825","1",148,0,4,2,818992
"20190826","2",149,0,1,3,833498
"20190827","3",152,0,0,0,772883
"20190828","4",149,1,0,0,824638
"20190829","5",143,1,0,0,868814
"20190830","6",149,2,2,2,792668
"20190831","7",149,0,3,2,830225
"20190901","1",152,0,4,1,841708
"20190902","2",156,0,0,0,829972
"20190903","3",153,0,0,0,893880
"20190904","4",153,0,0,0,821667
"20190905","5",152,0,0,0,909714
"20190906","6",152,0,0,0,869165
"20190907","7",152,0,0,0,709631
"20190908","1",152,0,0,0,775142
"20190909","2",152,0,0,0,867715
"20190910","3",152,0,0,0,791050
"20190911","4",152,0,0,0,820059
"20190912","5",159,0,0,0,814888
"20190913","6",79,0,0,0,850085
"20190914","7",157,0,0,0,791931
"20190915","1",156,0,9,2,749324
"20190916","2",154,0,0,0,853970
"20190917","3",157,0,0,0,798672
"20190918","4",159,0,0,0,820866
"20190919","5",158,0,0,0,847518
"20190920","6",160,0,0,0,898552
"20190921","7",161,0,0,0,806703
"20190922","1",158,0,9,1,803148
"20190923","2",165,0,0,0,828497
"20190924","3",163,0,0,0,958130
"20190925","4",162,0,0,0,911552
"20190926","5",160,0,0,0,827984
"20190927","6",156,0,0,0,811705
"20190928","7",160,0,0,0,845845
"20190929","1",160,0,9,1,841719
"20190930","2",160,0,0,0,811585
"20191001","3",162,0,0,0,817447
"20191002","4",162,0,0,0,990958
"20191003","5",160,0,0,0,880470
"20191004","6",48,0,0,0,910185
"20191005","7",158,0,0,0,900281
"20191006","1",157,0,9,2,824448
"20191007","2",158,0,0,0,1014439
"20191008","3",161,0,0,0,845978
"20191009","4",163,0,0,0,795588
"20191010","5",159,0,0,0,888827
"20191011","6",161,0,0,0,898451
"20191012","7",159,0,0,0,864003
"20191013","1",161,0,0,0,793733
"20191014","2",160,0,0,0,778911
"20191015","3",0,0,0,0,844110
"20191016","4",158,0,0,0,877697
"20191017","5",159,0,0,0,880386
"20191018","6",155,0,0,0,960184
"20191019","7",118,0,0,0,885271
"20191020","1",130,0,9,2,853794
"20191021","2",155,0,0,0,959543
"20191022","3",162,0,0,0,948837
"20191023","4",162,0,0,0,874080
"20191025","6",159,0,0,0,841795
"20191026","7",155,0,0,0,840083
"20191027","1",155,1,9,1,882195
"20191028","2",155,1,0,0,804998
"20191029","3",155,2,0,0,873209
"20191030","4",149,0,0,0,851670
"20191031","5",156,0,0,0,825803
"20191101","6",156,0,0,0,944468
"20191102","7",163,0,0,0,882644
"20191103","1",163,0,9,2,877474
"20191104","2",163,0,0,0,801663
"20191105","3",163,0,0,0,861581
"20191106","4",163,0,0,0,880810
"20191107","5",162,0,0,0,852350
"20191108","6",162,1,0,0,917474
"20191109","7",161,1,0,0,864701
"20191110","1",162,2,9,2,879234
"20191111","2",165,0,0,0,873829
"20191112","3",165,0,0,0,801106
"20191113","4",165,0,0,0,871859
"20191114","5",162,0,0,0,947029
"20191115","6",161,0,0,0,829755
"20191116","7",161,0,0,0,894351
"20191117","1",161,0,0,0,806946
"20191118","2",161,0,0,0,927961
"20191119","3",161,0,0,0,908957
"20191120","4",165,0,0,0,897570
"20191121","5",164,0,0,0,840253
"20191122","6",162,0,0,0,917120
"20191123","7",162,0,0,0,889958
"20191124","1",163,0,9,1,866761
"20191125","2",161,0,0,0,872199
"20191126","3",159,0,0,0,997972
"20191127","4",161,0,0,0,889243
"20191128","5",161,0,0,0,815678
"20191129","6",165,0,0,0,894917
答案 0 :(得分：0)
对于时间序列，如果您知道自己在做什么，最好使用ARIMA；如果不知道，最好使用先知;）