Bio inspired algorithms are machine learning algorithms which are modelled around observed natural phenomenon. Bio inspired algorithms in recent years have been shown to perform data mining tasks with ease by many research studies conducted for very complex domain of NP Hard problems. Ant Colony Optimizationis a bio inspired algorithmwhich addresses such complex problems which requires meta-heuristics based approaches to find solutions.

It is modelled to mimic movement of ants in search for food and is influenced by nature inspired computing paradigms. Computationally, Ant Colony Optimizationexplores and exploits solution space to find fairly accurate solutions in less time. Classification is a data mining function which assigns input values to two or more designated classesof output, with the goal of accurately predicting target class for new input data. Discriminant analysis which falls under the broad gambit of classification data mining, is the statistical mapping of data values to one of the two predefined groups. In this study, the application of Ant Colony Optimizationto perform discriminant analysis has been explored. Predictive classification accuracy ofeighty-six percent has been achieved on an ordinal data set for the combination of one thousand ants and twenty-five instances (minimum) covered by one rule. Comparative analysis of the quality of outcome based on factors like pheromone levels, number of ants and number of minimum instances covered by a rule has been modelled.

Authors: Hasnain Ali & Arpan Kumar Kar

Download the article here:

By Kar

Dr. Kar works in the interface of digital transformation and data science. Professionally a professor in one of the top B-Schools of Asia and an alumni of XLRI, he has extensive experience in teaching, training, consultancy and research in reputed institutes. He is a regular contributor of Business Fundas and a frequent author in research platforms. He is widely cited as a researcher. Note: The articles authored in this blog are his personal views and does not reflect that of his affiliations.