Baru-baru ini ditambahkan
-
Implementation of Cluster Analysis and Artificial Neural Networks as an Alternative for Klassen Typology and LQ: Case of Coconut
Due to the negative campaign supported by manufacturers of soy-oil in the United States that consuming coconut-oil may cause heart disease, coconuts' popularity continued to decline since the 60s. Coconut was abandoned by farmers with no attempt for rejuvenation for a long time. Coconut (Cocos nucifera) is widely distributed in tropical regions of Asia, Africa, Latin America and in the Pacific Islands. Economically, coconut is an important commodity in the Philippines, Indonesia, Papua New Guinea, Sri Lanka, South India, Malaysia, Tanzania and in the Caribbean. In the Philippines, as one of theworld's largest exporter of copra (dried coconut flesh) and coconut oil, more than 30 percent of the population is involved directly or indirectly to coconut industries for their lives [1]. Meanwhile, In Indonesia, coconut has long been known and greatly contributes to people's lives, either in terms of economic, social and cultural aspects [2]. Coconut trees are often referred to as the “tree of life” because from the coconut trees many products can be derived [1], from direct consumption as food and beverages to lubricant and surfactant for industrial purposes [3]. Other products that can be derived from coconuts are coconut sugar, Virgin Coconut Oil (VCO), and coconut water. Recent research shows that coconut products are good for health and promoted as a superfood. Withthe help of celebrities, coconut is dubbed as a superfood, result in a surge in demand. The increasing demand mainly comes from the coconut water and VCO [2]. The global market is now witnessing the demand for coconut products increase significantly. Globally, demand for coconut products growing exponentially [4]. With the current trends, the estimated global market for coconut will increase to USD 10 billion by 2030. This increased demand for coconut products mostly originated from the higher added value products which in turn will provide opportunities to increase the income for millions of smallscale coconut producers [4]. Although there is optimism or increasing coconut demand in the future, the previous prediction from the World Bank in the 80s stated that coconut oil compared to the main competitor did not give bode well for the communities who rely on coconut as the source of main income [1]. The latest data show that the share of coconut oil and its growth relatively minuscule [5]. Since the price of coconut products are still volatile, mixed cropping between coconut and the perennial plant may improve and stabilize the farmers’ income. For instance, coconut can be mixed with cacao or coffee. The aim of the research is implementing clustering technique, ie. Dendogram, Principle Component Analysis (PCA) and Kohonen Self Organizing Map (Kohonen SOM) in place where Klassen Typology and Location Quotient (LQ) are commonly used to identify the base (leading) commodities in certain area. Klassen Typology or LQ can be used to classify the potential area to be developed for coconut. Nevertheless, these approaches are not suitable to be implemented for cases with large data. Klassen Typology and LQ are also difficult to be combined with other data, such as socio-economic data. They only portrait the comparative condition for the area being analyzed. Since LQ and Klassen Typology are categorization processes, this task can also be handled by cluster analysis. This paper uses cluster analysis, either based on a statistical approach or an artificial neural network. Both can be classified as machine learning. Dendrogram, PCA and Kohonen SOM were implemented to identify the coconut industry at the provincial l vel along with other commodities. Cluster analysis, especially Kohonen SOM has many flexibilities. For instance, data from other sectors can be easily added to the model. They are also able to handle large data by learning the pattern of the data provided and make a generalization.