This paper reviews the different information retrieval models that can be used to represent information from digital library. It also describes document retrieval from the digital library, followed by experimental analysis of various traditional IR models like Boolean model, Vector model, Probabilistic model on a sample corpus. We present the Private Digital Library (PDL) project that represents a service of the Corporate Digital Library (CDL) prototype. The main ideas underlying this project are the following. When a user is looking for documents that he already retrieved in the past, he has to repeat the search procedure and solve the same problems he encountered in the past access. On the contrary, consider the possibility for the user to store documents in a private library. In this case, he will have the chance to retrieve the documents of interest more easily and quickly. It may also be the case that the user has not a clear understanding of what he is looking for. Moreover, he might not know exactly the library content and organization. This paper also discusses way of representing index terms in form of indices so that they can be searched later. Thus, he needs for assistance to suggest him what to look for and how to query the system. A method to store and catalogue documents according to personal criteria might help to overcome these problems.

Key words : ,,