Avtomatsko opisovanje vsebine dokumentov na internetu
Author is drawing attention to the width of the conception of the subject of Internet's information objects. The conception comprises the subject of texts and non-textual documents as well as the subject of information needs. Due to the size and dynamic nature of the Internet the construction of the large Web-based document databases call for automatic selection of data about documents, including the data about documents' subject. The procedures used are shortly described in this article. Automatic subject description is especially difficult in cases when we need the description elements to be of the same data type as the documents, e.g. images described with the image primitives.