Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 205
194 I N N O VAT I O N S I N T R AV E L D E M A N D M O D E L I N G , V O L U M E 2 (a) (b) FIGURE 6 Measurement results of changes in acceleration at (a) coffee shop and (b) record shop. ies, which allow calculation only with paper and pencil, have been considered superior in such cases. However, expectations are that enormous amounts of travel behav- ior data will continue to be stocked in databases as a result of progress in survey techniques, which has been shown in this paper. In that case, an effective analysis method will be data mining, which directly mines strate- gies effective for transportation policies from a large amount of data, unlike conventional approaches, which try to validate assumptions and reproducibility of mod- els by using data. A well-known application of data mining is the diapers-and-beer episode of Wal-Mart, the largest U.S. FIGURE 7 Ranges in atmospheric pressure. retailer. A correlation rule analysis, a typical method of data mining, was performed on an enormous amount of pheric pressure sensors are effective for identifying purchase data, which were being instantly collected indoor floors that are outside the range of GPS radio through a point-of-sale system, and it revealed a rule that waves. It also seems possible to construct, from such customers buying diapers on Friday evening tend to buy information, an automatic estimation model for behav- cans of beer as well. Wal-Mart immediately placed cans ioral contexts without forcing subjects to perform any of beer beside the diaper section, and beer sales doubled. action, by using a hidden Markov-type model. This episode indicates that it is possible to directly draw causal relationships of consumer behaviors, which are difficult to obtain by intuition of analyzers, from a DATA-ORIENTED APPROACHES large amount of data. This method is already in use for various business data analyses, such as those for inven- Such techniques that enable long-term online observa- tory control, new product planning, securities valuation, tions of travel-activity patterns may also have a great stock valuation, and medical diagnosis, and it is deliver- influence on the usage of behavior models. Such moni- ing remarkable results. toring techniques will significantly affect the design of Unlike conventional analysis approaches, which have fares for public transportation, central urban area plan- poor flexibility in policy evaluation because of too much ning, and transportation planning in real time. It is emphasis on consistency and reproducibility of models, because the behavior databases, in which data continue data mining captures transportation policies in a mar- to be accumulated through IT-based monitoring tech- keting sense, on the basis of a large amount of data, and niques, themselves will contain travel behavior models focuses on discovering effective relationships from the and activity models, making possible simultaneous data. searches for and viewing the findings on various travel In fact, such a method based on data science is not behaviors. considerably different from conventional approaches, Both techniques for measuring travel behavior and which consist of the steps of data sampling, analysis, computer techniques are ceaselessly progressing. It is modeling, validation, and then solution of real problems. impossible to create a program on a computer when the The only difference is that such a method tries to make object to be calculated is not clear, and theoretical stud- data themselves reveal many relationships. Such a