Using normal distribution to retrieve temporal associations by Euclidean distance

Euclidean distance measure is widely adopted distance measure to find the distance between any two vectors. In this paper, we extend the use of Euclidean distance to the context of normal distribution based temporal pattern mining. The similarity between any two patterns is computed by using the probability vectors of corresponding temporal patterns. These temporal patterns in our case are expressed as probability vector sequences. The probabilities are found from computed normal scores of patterns considering a given reference and using probability chart. The work in this paper is restricted to introducing the approach of mining patterns using the proposed dissimilarity measure.