SUBJECT: Re : Hierarchical Corpuses OHSUMED is technically not a hierarchical collection of documents . It is a flat collection of documents ( bibliographic records , to be precise ) that is indexed by a hierarchical vocabulary , MeSH . &NAME <QUOTE> <QUOTE> <QUOTE> &EMAIL &NUM / &NUM / &NUM &NUM : 43AM ) ) ) Hi &NAME , <END_QUOTE> <END_QUOTE> <END_QUOTE> You should consider OHSUMED which is a set of medical abstracts that have been manually classified using MeSH ( Medial Subject Headings ) . I have published some results on this corpus in the following paper . Hierarchical text categorization using neural networks &NAME &NAME & &NAME &NAME Information Retrieval , &NUM ( &NUM ) &NUM . January &NUM Let me know if you would be interested in having a copy of it . &NAME &NAME On Thursday &NUM April &NUM &NUM : &NUM am , &NAME &NAME wrote : <QUOTE> &NAME , I am looking for a large hierarchical corpus of text documents , for a research in the field of hierarchical classification . I would like to get your opinion about the followings : &NUM &NAME - &NAME Collection Version &NUM . Is it 's hierarchy deep enough to make a point when it comes to showing the dependency of classification results on the hierarchical structure of the corpus ? &NUM &NAME - Open directory project . Why are there no researchs refering to data from the &NAME ? &NUM &NAME ! - There are several works done on data which was gathered from &ORG's science hierarchy . How can I get a copy of it ? Do I need &ORG's permission ? How do I get it ? &NUM US Patents - Where can I get this one ? &NUM Any other suggestions ? Since I could not find many research works using the &NAME corpus , even though it looks like a good one for a comparison bassis among works in the field , and several works are using &ORG's science hierarchy , it seems like a good idea to have both tested and put into the research . Looking Fwd to getting your comments on those thoughts , &NAME &NAME &NAME University . <END_QUOTE> Dr. &NAME &NAME &NAME School of Informatics Department of Library and Information Studies &NUM &NAME &NAME &NAME , NY &NUM &NAME : ( &NUM ) &NUM ext . &NUM &NAME : ( &NUM ) &NUM