I want to use LIBSVM for electrical load forecasting but my prepared data stucked. People says that labels must be any real number but my data electrical load. How can i convert data or handle this issue?
My Libsvm format: Labels represent electrical load, 1 represents days(1Monday..5 Friday), 2 represents hours(0...23), 3 represents temperature, 4 represents previous load. I have 1745 row like these.I used firslty svm-train.exe( approximtely 70% of data) after svm-predict.exe. Accuracy comes 0 every time. I tried scaling datas. What is the problem in this format? Where am I doing wrong? If help me, I will be very appreciated.
147.306 1:5 2:0 3:-5 4:125.696
125.696 1:5 2:1 3:-5 4:113.152
113.152 1:5 2:2 3:-4 4:108.288
108.288 1:5 2:3 3:-3.6 4:98.816
98.816 1:5 2:4 3:-3.3 4:99.328
99.328 1:5 2:5 3:-3.6 4:110.08
110.08 1:5 2:6 3:-2.6 4:102.912
102.912 1:5 2:7 3:-2.6 4:118.784
