Binary Information Press
Sequences in restaurant recommendation system meet two constraints: begin with entry-point and end in end-point. For effectiveness and efficiency considerations, constraints are essential in mining patterns on this system. In this paper, two discovered methods of constraint based patterns and rules are proposed which generate more efficient results. Performances on sequences of restaurant recommendation system have demonstrated that the proposed algorithms could significantly reduce the runtime and generate compact patterns in the process of mining maximal and constraint based patterns, as well as generate effective sequential rules.