I've got about 10000 txt files. Every txt file contains video metadata in form:
Title: ...\n
Video Id: ...\n
Url: ...\n
Duration: ...\n
and other attributes
I want to cluster these videos using their metadata with k means in weka, but I have problem preprocessing them. I load them with textDirectoryLoader, but I want the attributes to be Title, Photo Id, Url, Duration etc. Is there any way to bring data in this form in the arff file?