An Auto-Associative Neural Network for Information Retrieval

Neural network is an important paradigm that has received little attention from the community of researchers in information retrieval, especially the auto-associative neural networks. These networks are capable of discovering patterns of terms among documents. We propose an auto-associative neural network to model the classification and to perform the matching task The unique layer network is trained with the documents of the collection and then used to recall the most relevant documents to specific queries. Our model has been tested on a TREC sub-collection. The results are compared against the vector space model, The experiment shows higher level of global precision and recall. The recall-precision curves show an important improvement on the precisions for the low levels of recall, which indicates a faster retrieval of the first relevant documents. This strength of the auto-associative neural network makes it an attractive model in information retrieval for general collections.