Learning decision trees using parallel sequential induction network

In a decision making (or pattern classification) process, many factors (or attributes of the pattern) are evaluated. These attributes may have different ranking of importance in discriminating patterns against each other. Some of them may be independent and can be evaluated at the same time, i.e. in parallel. Others may have to wait until a preliminary decision (or classification) based on some more important attributes has already been made. These combination of both the parallel and sequential strategy forms the basis of the ''Parallel Sequential Induction Network'' (PSIN) for decision making process.