Predict Future Events Over Smart Environment Through Modified Apriori Algorithm

In a smart environment, mining of patterns of occurrences of the past events leads to the prediction of the appropriate event to be occurred as next event. As such, this work is to identify the challenges appearing in smart environment research with respect to finding the pattern of occurrences, followed by a proposed solution. The proposed solution is designed and implemented over the existing Apriori data mining algorithm with certain required modifications to incorporate the time information of occurrences of events. During the implementation of the said modified algorithm, appropriate datasets related to the events of daily activities of the inhabitants of the given environments have been provided; and it is found that the implementation has given the appropriate results with respect to the requirements of the work. Moreover, one more modification over the Apriori algorithm is made for considering all possible subsets of a candidate item set; otherwise impact of the introduction of the time duration information is not up to the level as desired. From the experiment through implementation, it is claimed that the proposed mechanism can choose the next event from the list of predicted events more accurately after inclusion of time duration information between consecutive events.