Succinct Indexes for Reporting Discriminating and Generic Words

We consider the problem of indexing a collection $\cal{D}$ of D strings (documents) of total n characters from an alphabet set of size σ, such that whenever a pattern P (of p characters) and an integer i¾? ∈ [1, D] comes as a query, we can efficiently report all (i) maximal generic words and (ii) minimal discriminating words as defined below: These problems were introduced by Kucherov et al.a[8], and they proposed linear space indexes occupying O(nlogn) bits with query times O(p + output) and O(p + loglogn + output) for Problem (i) and Problem (ii) respectively. In this paper, we describe succinct indexes of nlogσ + o(nlogσ) + O(n) bits space with near optimal query times i.e., O(p + loglogn + output) for both these problems.