探討數據整合之方案相似與數據尺度

Investigating the Alternative Similarity and Scale Variation in Data Combination

研究目的在於探討數據整合模式中方案相似與數據尺度之影響。在過去運輸文獻中，巢式羅吉特架構雖能整合多重數據，但其架構未能考慮不同方案間的數據尺度。因此，本研究首先運用誤差成分羅吉特（Error Components Logit,ECL）模式進行數據整合，針對數據尺度與方案相似進行探討。實證結果顯示，ECL模式除可考量不同運具間的數據尺度變異外，亦可有效衡量運具間的方案相似性。在數據尺度的結果中，敘述性偏好的數據變異程度高於顯示性偏好，且私人運具高於大眾運輸。在方案相似性方面，可分成機車與汽車、公共巴士與輕軌兩個競爭集群。

This study aimed to investigate the influences of alternative similarity and scale variation in combining multiple data sets. The structure of nested logit was frequently used to merge revealed preference and stated preference data sets in previous literatures. Howeνer, it could not identify the scale variation across mode alternatives. Hence, this study proposed error components logit model to combine multiple data sets and to explore the effects of alternative similarity and scale variation. Empirical results revealed that the proposed model could not only identify the scale difference across modes, but also explore alternative similarity among modes. The result of data scale indicates that the variation of stated preference is higher than revealed preference and the νariation of private transportation is higher than public transportation. Furthermore the choice set of mode alternatiνes could be segmented into two competitive groups: private mode (motorcycle and car) and public mode (bus and light rail transit).