A Systematic Literature Review of the effect of choice of a distance/similarity/dissimilarity metric on the performance evaluation of a clustering/classification algorithm

This work was done as a research module (in my MSc) to demonstrate research writing abilities and explore a potential gap in the existing research, the context of which is mainly reviewing the effect of the choice of a distance/similarity/dissimilarity measure on clustering/classification algorithm performance. The work follows the guidelines for performing Systematic Literature Reviews in Software Engineering proposed by Kitchenham.

The results of this review indicate that the effect of the choice of a distance/similarity/dissimilarity metric on the performance of the clustering/classification algorithm is dependent on the type of data and the task being dealt with. Also, Domain-based modified (or derived or anomalous) distance/similarity/dissimilarity metrics produce better clustering/classification results.

Future work will involve undertaking a trial or experiment.

Please find attached each deliverable and its feedback on the work.

Original link (bitbucket): https://bitbucket.org/bilal658/7com1085-group93/src/master/