臺灣省政府資訊資源預算規劃及效益評估之研究

Abstract

推行業務電腦化是臺灣省政府近年來的生點工作之一，面對每年龐大的資訊投資金額，省府資訊作業主管機關面臨了如何客觀「分配資訊預算」及「進行資訊作業效益評核」二項重大課題。故本研究的目的即在嘗試結合「模糊綜合評判（Fuzzy Multicriteria Decision Making）」與「類神經網路（Neural Network）」二種模式的特性，建立一「省府資訊預算規劃與效益評核輔助系統」的架構。其中將利用所定義的各項主、客觀考慮因素及指標進行全面性資訊作業效益評核；並以資訊作業績效顯著機關的歷年行為模式作為未來資訊預算規劃的準則，依此對省屬各機關作適當的行為監控。 本研究實際收集了省府部分機關之七十七至八十二年度的相關資料，並依系統架構進行各項分析。分析的結果中，在資訊作業效益評核方面，各類機關（行政、事業及金融保險等三類）的平均機關誤差比分別僅為43.48%、53.33%及40%。而在資訊預算審查方面，先以類神經網路掌握78至81年度的行為模式，再以82年度資料來驗證，初步而言，已能掌握大致的行為模式，但是仍有待更多年度資料的累積，以使行為模式的掌控更為精確。

To computerize operations is one of the major tasks of the Provincial Government of Taiwan. When making a large quantity of investment in information resources, it has to deal with two problems-how to plan information resources budgets and how to evaluate the performance objectively. The primary goal of this study is to combine the fuzzy multicriteria decision making method with the neural network model to propose a structure of computer-aided budget planning and performance evaluating system for the Provincial Government of Taiwan. This system uses defined subjective factors and objective pointers to evaluate the performance of each department; taking the historical behavior of departments with better performance as the guidelines of budget planning. On-site data of some departments of the Provincial Government of Taiwan for the period 1988 was 1993 was collected to do the two analysis. In the performance evaluation analysis, the average department difference rate of executive, enterprise and financial division is 43.48%, 53.33% and 40%. In information resources budget planning analysis, data from 1989 to 1992 was used to train the network and then input 1993’s data to test it. In the result, we find the network has grasped mainly the behavioral mode of budget planning. But more data is needed to train the network to get a more precise result.