Computation Analysis for Finding Co–Location Patterns using Map–Reduce Framework

Abstract

Objectives: The main objectives of the paper are 1. Generating the Neib_tree based on the number of features and instances. 2. Finding all the co-location patterns using Parallel Approach. 3. Improving the computation time by exploring Map-Reduce Framework. Methods: To generate Neib-tree is by Grid based approach. The method used find co-location patterns is by parallel approach which drastically increases the time complexity. The exploratory results are directed by utilizing manufactured information sets by taking the different data sets one with 25k, 50k and 75k features and an average of 20k instances each, which produces the computational analysis with a distance of 20km. Findings: This paper presents fast calculation of co-location patterns where these is helpful in finding the people suffering from a particular problem in a place and what are the patterns affecting the problem. The proposed work diminishes the calculation time by 1/n terms where n is the quantity of components as it uses a Map-Reduce system. This paper presents exact and fulfillment of the new approach. At long last, exploratory assessments utilizing manufactured information sets demonstrate the calculation is computationally more productive. Applications: The concept presented in this paper is helpful in different areas like medical Field, NASA, and etc., Improvements: The paper improves the time complexity and space complexity by using parallel join-less approach.