A type of automatic indexing based on term weighting and relevance prediction that attempts to predict probable query search terms based on term frequencies within documents and the inverse document frequency of terms across the target data. It is similar to the probabilistic model