Developing Agriculture Purchasing Managers’ Index for Describing Taiwan’s Agriculture Industry by Using Automatic Weighted K-Means Algorithm

Although the average agricultural total output is not greater than 2% of Taiwan’s GDP, the Taiwan government still attaches great importance to the agriculture industry to ensure the food self-sufficiency rate. Taiwan still has no indicators to measure the status of the agriculture industry. This paper proposes an idea to develop Agriculture Purchasing Managers’ Index (APMI) for Taiwan agriculture industries. To reduce the effect of some statistical assumptions and to provide more clarity and direct analysis of results, this paper proposes a novel automatic weighted k-means algorithm to develop the APMI. The results of this research suggest that four variables should be included in the APMI of the pig industry, namely “Trade amount”, “Average weight per pig”, “Price per pig”, and “Slaughtered.”. Among these, “Slaughtered” and “Trade amount” are the more important variables for developing the APMI of the pig industry. The proposed model offers three advantages: (a) it can be successfully used to construct APMI, (b) It can automatically search the weight of each variable without any human judgment in APMI, and (c) It avoids some statistical assumptions and explains the results more clearly and directly. Thus, the proposed model can be used to construct used APMI proposed in this work, and it describes the status of the agriculture industry.

Keywords: pig industry; trade of pigs; automatic weighted selection

*Corresponding author: Tel.: (+886) 7-8100888 # 25322

                                             E-mail: cplee@nkust.edu.tw