Contextual probability (G) provides an alternative, efficient way of estimating (primary) probability (P) in a principled way. G is defined in terms of P in a combinatorial way, and they have a simple linear relationship. Consequently, if one is known, the other can be calculated. It turns out G can be estimated based on a set of data samples through a simple process called neighbourhood counting. Many results about contextual probability are obtained based on the assumption that the event space is the power set of the sample space. However, the real world is usually not the case. For example, in a multidimensional sample space, the event space is typically the set of hyper tuples which is much smaller than the power set. In this paper, we generalise contextual probability to multidimensional sample space where the attributes may be categorical or numerical. We present results about the normalisation constant, the relationship between G and P and the neighbourhood counting process.
Probability estimation Contextual probability Neighbourhood counting
This is a preview of subscription content, log in to check access.
Ash, R.B., Doléans-Dade, C.: Probability and Measure Theory. Academic Press, San Diego (2000)zbMATHGoogle Scholar
Duda, R.O., Hart, P.E.: Pattern Classification and Scene Analysis. Wiley, New York (1973)zbMATHGoogle Scholar
Feller, W.: An Introduction to Probability Theory and Its Applications. Wiley, New York (1968)zbMATHGoogle Scholar
Hajek, A.: Probability, logic and probability logic. In: Goble, L. (ed.) Blackwell Companion to Logic, pp. 362–384. Blackwell, Oxford (2000)Google Scholar
Lin, Z., Lyu, M., King, I.: Matchsim: a novel similarity measure based on maximum neighborhood matching. Knowl. Inf. Syst. 32, 141–166 (2012)CrossRefGoogle Scholar
Mani, A.: Comparing dependencies in probability theory and general rough sets: Part-a. arXiv:1804.02322v1
Mani, A.: Probabilities, dependence and rough membership functions. Int. J. Comput. Appl. 39, 17–35 (2017)Google Scholar
TolgaKahraman, H.: A novel and powerful hybrid classifier method: development and testing of heuristic k-nn algorithm with fuzzy distance metric. Data Knowl. Eng. 103, 44–59 (2016)CrossRefGoogle Scholar