Studentski izvještaji

Izrađena je CSV datoteka s podacima za učenje.

Primjer datoteke:

0.020750,0.397349,0.093680,0.095192,0.091558,0.095130,0.070817,0.808820,0.323488,0.585612,0.013106,0.032234,0.211226,JPG
0.036282,0.131566,0.058351,0.001613,0.007259,0.003279,0.168669,0.959480,0.119376,0.125695,0.032234,0.062128,0.093143,JPG
0.013559,0.126013,0.080064,0.274000,0.231599,0.039639,0.128541,0.676648,0.041617,0.012520,0.020784,0.015086,0.925766,JPG

Redci te datoteke su rezultati izvođenja testova na blokovima od 4096 bajta podataka. Prvi blok svake datoteke se preskače (header). Stupci te datoteke su 13 testova (float brojevi) te poseban stupac koji označava kategoriju.

Ta datoteka se učita i njome se istrenira klasifikator K najbližih susjeda. Nakon toga je klasifikator spreman za klasificiranje nepoznatih podataka.

Ocjenu rada klasifikatora možemo dobiti i iz samih podataka korištenih za učenje.

Klasifikator je pokazao da su mu točnosti sljedeće:

JPG: 0.78
MP3: 0.83
TXT: 0.93
RAR: 0.80
PDF: 0.74
PNG: 0.86