ABSTRACT
This study aims to analyze a dataset of plants and configure the most suitable plant species for a particular area, hoping that research in this area will elevate the planting process by reducing time and effort that goes in when deciding for the suitable plant type for a given area. The method carried out for analysis is by using a dataset that consists of 104,352 plant species planted in the city of Greater Geelong council region, Australia. It was analyzed through multivariate graphs, clustering and classification algorithms in machine learning technology. Results clearly show that by using K-means algorithm, Citrinus (Callistemon Citrinus) is the most suitable plant for the area of Geelong, Australia, and various graphical analyses carried out using health status, maturity and structure of the species verified also the findings. The p-value was determined using one-way ANOVA method and it is 0.014 between the area clusters (p=0.014) showing the significance among areas. Additionally, J48 classification algorithm gives the best result for classification of plant species. However, all the algorithms also showed good results as the correctly classified percentage was approximately 90% with only 2% of fluctuation among algorithm result criteria. This inquiry has found that the K-means algorithm can be used with a dataset including species, location, height, crown width, DBH (Diameter at Breast Height) and health status in order to determine the suitable plant. Also, J48 is one of the best prediction methods having a high percentage of correctly classified instances. Furthermore, analyzing plant species against their growth and health will help to understand and reflect on declining plant species in areas. The study can be taken as a guide when selecting a plant species to grow in a specific location for agricultural, gardening or any other industrial plantation. Although this research was conducted in the area of the city of Greater Geelong council, Australia, it can be applied to any similar temperate climates.
