模糊統計分類在臺灣地區失業率分析與預測之應用

Abstract

失業率是各國經濟發展所應重視的指標之一。過高的失業率會影響整個就業市場 及社會秩序。近年來，由於台灣工資高漲，加上大量引進外勞而導致失業率升高，已造成社 會的嚴重問題。此一問題若未獲得改善，不僅形成人力資源的閒置，且將造成嚴重的社會失 調。 本文即以失業率代表勞動市場供需變數之指標，嘗試以模糊時間數列分類與辨識的方 法，應用平均累加模糊熵〈average of the sum of fuzzy entropies〉，來認定台灣地區失 業率走勢與結構改變。由實證結果發現，應用模糊統計分析方法比傳統中的測度方法能得到 更合理的解釋，且預測結果可以提供決策者更多的資訊，做出正確的決策。

The percentage of unemployment is one of the targets that every country emphasizes in the social-economic development. The soaring percentage of lasting unemployment would sure]y influence the wholesome society. Taiwan has been making an amazing progress in economy and thus are ranked as one of the well-developed country. The imblanced development of the economy, however, has caused the problem of the sources of manpower. The increasing rise of unemployment manifests the imblance of the allocation of manpower. This paper uses the percentage of unemployment as the indication of the supply-and-need variable in labor market. Applying the fuzzy time series classification method, we make an analysis of the variations in Taiwan labor market and seek a further solution to lower the percentage of unemployment. Seeing that the factors to evaluate are prone to change, we should take the fuzzy weight into account. Finally, instead of the application of traditional methods of measuring, fuzzy statistical classification is expected to reflect the current circumstances.