台灣各縣市失業率之灰分析

A Grey Analysis of Regional Unemployment in Taiwan

自2000年以來，台灣失業率屢創新高，失業問題已成為各界所關注的焦點。本研究先以「灰關聯分析法」找出與失業率灰關聯度較高之因子，再以台灣地區及台北市為例，透過這些因子進行失業率之「灰預測」，並根據預測結果評估「灰色理論」在失業議題上的適用性。本文所採取的是各縣市1998年至2006年主計處統計資料，研究結果發現45-64歲人口比例、服務業人口比例、高等教育人口比例、營利事業家數與台灣地區失業率的灰關聯度較高，農林漁牧業人口比例的灰關聯度最低，另亦得出各縣市之灰關聯序。另外，透過灰關聯序進行變數篩選後，可得準確度相當高之失業率灰預測模型，其中，台灣地區失業率預測準確度高達99.58%，台北市之失業率預測準確度更高達99.72%，可見灰色理論可適用於區域失業議題的分析與預測上。

Much attention has been paid to the high and rising unemployment level in Taiwan since 2000. This paper uses the grey relational analysis (GRA) to compute the grey relational ordinal (GRO) for unemployment-rate-related regional factors and then constructs the appropriate grey models (GM) based on the ordinal to forecast unemployment rates for Taiwan and Taipei County. The forecasting accuracy is examined to judge whether or not the grey theory is appropriate for studying the unemployment issue. All of the data at the county level come from the Directorate General of Budget, Accounting and Statistics (DGBAS), from 1998 to 2006. Our major findings are as follows: The proportion of age 45-64, proportion of service industry employees, proportion of higher education attainment, and number of business units are highly grey correlated with the unemployment rate in Taiwan. The employment proportion of agriculture, forestry and fisheries in contrast has the lowest grey correlation grade with the unemployment rate. We also acquire the GRO of 23 counties. Once eliminating factors on the basis of GRO, we can obtain the best grey model providing excellent forecasting ability. The forecasting accuracy is 99.58% for the unemployment rate of Taiwan and 99.72% for Taipei County. The grey approach has shown its power in analyzing and forecasting regional unemployment issues, as demonstrated by this study.