Bayesian Network Analysis for the Questionnaire Investigation on the Impression at Yoshiwara Shopping Street in Fuji City

Shopping streets at local city in Japan became old and are generally declining. In this paper, we handle the area rebirth and/or regional revitalization of shopping street. We focus on Fuji city in Japan. Four big festivals are held at Fuji city. Many people visit these festivals including residents in that area. Therefore a questionnaire investigation to the residents and visitors is conducted during these periods in order to clarify residents and visitors’ needs for the shopping street, and utilize them to the plan building of the area rebirth and/or regional revitalization of shopping street. There is a big difference between Fuji Shopping Street and Yoshiwara Shopping Street. Therefore we focus Yoshiwara Shopping Street in this paper. These are analyzed by using Bayesian Network. Sensitivity analysis is also conducted. As there are so many items, we focus on “The image of the surrounding area at this shopping street” and pick up former half and make sensitivity analysis in this paper. The analysis utilizing Bayesian Network enabled us to visualize the causal relationship among items. Furthermore, sensitivity analysis brought us estimating and predicting the prospective visitors. Sensitivity analysis is performed by back propagation method. These are utilized for constructing a much more effective and useful plan building. We have obtained fruitful results. To confirm the findings by utilizing the new consecutive visiting records would be the future works to be investigated.


INTRODUCTION
Shopping streets at local city in Japan are generally declining.It is because most of them were built in the so-called "High Growth Period (1954Period ( -1973))".Therefore they became old and area rebirth and/or regional revitalization are required everywhere.
There are many papers published concerning area rebirth or regional revitalization.Inoue (2017) has pointed out the importance of tourism promotion.Ingu et al.(2017) developed the project of shutter art to Wakkanai Chuo shopping street in Hokkaido, Japan.Ohkubo (2017) has made a questionnaire research at Jigenji shopping street in Kagoshima Prefecture, Japan and analyzed the current condition and future issues.For about tourism, many papers are presented from many aspects as follows.
Yoshida et al. designed and conducted a visitor survey on the spot, which used a questionnaire to investigate the activities of visitors to the Ueno district in Taito ward, Tokyo.Doi et al. analyzed the image of the Izu Peninsula as a tourist destination in their 2003 study "Questionnaire Survey on the Izu Peninsula."Kano conducted tourist behavior studies in Atami city in 2008, 2009, 2014 and in other years.In this paper, we handle the area rebirth and/or regional revitalization of shopping street.We focus on Fuji city in Japan.Fuji city is located in Shizuoka Prefecture.Mt.Fuji is very famous all around the world and we can see its beautiful scenery from Fuji city, which is at the foot of Mt.Fuji.There are two big shopping street in Fuji city.One is Yoshiwara Shopping Street and another one is Fuji Shopping Street.They became old and building area rebirth and regional revitalization plan have started.Following investigation was conducted by the joint research group (Fuji Chamber of Commerce & Industry, Fujisan Area Management Company, Katsumata Maruyama Architects, Kougakuin University and Tokoha University).The main project activities are as follows.
A. Investigation on the assets which are not in active use B. Questionnaire Investigation to Entrepreneur C. Questionnaire Investigation to the residents and visitors After that, area rebirth and regional revitalization plan were built.
In this paper, we handle above stated C.
Four big festivals are held at Fuji city.Two big festivals are held at Yoshiwara Shopping Street and two big festivals at Fuji Shopping Street.
At Yoshiwara Shopping Street, Yoshiwara Gion Festival is carried out during June and Yoshiwara Shukuba (post-town) Festival is held during October.On the other hand, Kinoene Summer Festival is conducted during August and Kinoene Autumn Festival is performed during October at Fuji Shopping Street.Many people visit these festivals including residents in that area.
Therefore questionnaire investigation of C is conducted during these periods.
Basic statistical analysis and Bayesian Network analysis are executed based on that.
In this paper, a questionnaire investigation is executed in order to clarify residents and visitors' needs for the shopping street, and utilize them to the plan building of the area rebirth and/or regional revitalization of shopping street.There is a big difference between Fuji Shopping Street and Yoshiwara Shopping Street.Therefore we focus Yoshiwara Shopping Street in this paper.These are analyzed by using Bayesian Network.Sensitivity analysis is also conducted.As there are so many items, we focus on "The image of the surrounding area at this shopping street" and pick up former half and make sensitivity analysis in this paper.By that model, the causal relationship is sequentially chained by the characteristics of visitors, the purpose of visiting and the image of the surrounding area at this shopping street.The analysis utilizing Bayesian Network enabled us to visualize the causal relationship among items.Furthermore, sensitivity analysis brought us estimating and predicting the prospective visitors.Sensitivity analysis was conducted by back propagation method.

OUTLINE AND THE BASIC STATISTICAL RESULTS OF THE QUESTIONNAIRE RESEARCH Outline of the Questionnaire Research
A questionnaire investigation to the residents and visitors is conducted during these periods in order to clarify residents and visitors' needs for the shopping street, and utilize them to the plan building of the area rebirth and/or regional revitalization of shopping street.

Basic Statistical Results
Now, we show the main summary results by single variable.

Summary results for the items used in Hypothesis Testin
(1) How often do you come to this shopping street?(Q1) Everyday 12.9%,More than 1 time a week 15.6%,More than 1 time a month 23.4%,More than 1 time a year 37.3%, First time 5.1%,Not filled in 5.6% These are exhibited in Figure 4. (2) What is the purpose of visiting here?(Q2) Shopping 20.7%, Eating and drinking 13.1%, Business 7.5%, Celebration, event 47.5%, Leisure, amusement 1.5%, miscellaneous 9.7% These are exhibited in Figure 5. (3) How do you feel about the image of the surrounding area at this shopping street?(Q3) Beautiful 51.9%, Ugly 48.1%,Of the united feeling there is 47.We used BAYONET software (http://www.msi.co.jp/BAYONET/).When plural nodes exist in the same group, it occurs that causal relationship is hard to set a priori.In that case, BAYONET system set the sequence automatically utilizing AIC standard.Node and parameter of Figure 8 are exhibited in Table 1.In the next section, sensitivity analysis is achieved by back propagation method.Back propagation method is conducted in the following method (Figure 9).

SENSITIVITY ANALYSIS
Now, posterior probability is calculated by setting evidence as, for example, 1.0.Comparing Prior probability and Posterior probability, we can seek the change and confirm the preference or image of the surrounding area at this shopping street.We set evidence to all parameters.Therefore, the analysis volume becomes too large.In this paper, we focus on "The image of the surrounding area at this shopping street" and pick up latter half and make sensitivity analysis.We prepare another paper for the rest of them.
As stated above, we set evidence for each parameter, and the calculated posterior probability is exhibited in Appendix 2. The value of "Posterior probability -Prior probability" (we call this "Difference of probability" hereafter) is exhibited in Appendix 3. The sensitivity analysis is executed by mainly using this table.
Here, we classify each item by the strength of the difference of probability.
Strong (++, ): Select major parameter of which absolute value of difference of probability is more than 0.05 Medium (+, ): Select major parameter of which absolute value of difference of probability is more than 0.01 From To From To We can observe that "Those who have an image of the surrounding area at this shopping street as "Fascinating" had come under the image of the surrounding area at this shopping street as "Beautiful", "Of the united feeling there is", "Varied"," Cheerful", "Individualistic", "Friendly", "Healed", "Want to play" or "Lively" of an age of "10th" or "20th" in which the gender is "Female".
(10) Setting evidence to "Not fascinating" After setting evidence to "Not fascinating", the result is exhibited in Table 11.We can observe that "Those who have an image of the surrounding area at this shopping street as "Not fascinating" had come under the image of the surrounding area at this shopping street as "Featureless", "Urban", "Gloomy" or "Want to examine deliberately" of an age of an age of "30th", "60th" or "More than 70" in which the gender is "Female".
(11) Setting evidence to "Want to play" After setting evidence to "Want to play", the result is exhibited in Table 12.We can observe that "Those who have an image of the surrounding area at this shopping street as "Want to play" had come under the image of the surrounding area at this shopping street as "Of the united feeling there is", "Cheerful", "Individualistic", "Friendly", "Healed", "Fascinating", "Lively" or "Atmosphere of urban" of an age of "10th", "20th" or "40th" in which the gender is "Male".
(12) Setting evidence to "Want to examine deliberately" After setting evidence to "Want to examine deliberately", the result is exhibited in Table 13.We can observe that "Those who have an image of the surrounding area at this shopping street as "Want to examine deliberately" had come with the purpose of visiting for "Shopping" under the image of the surrounding area at this shopping street as "Ugly", "Gloomy" or "Not fascinating" of an age of "40th", "60th" or "More than 70" in which the gender is "Female".
(13) Setting evidence to "Lively" After setting evidence to "Lively", the result is exhibited in Table 14.We can observe that "Those who have an image of the surrounding area at this shopping street as "Lively" had come under the image of the surrounding area at this shopping street as "Beautiful", "Of the united feeling there is", "Varied", "Cheerful", "Individualistic", "Friendly", "Healed", "Fascinating", "Want to play", or "Atmosphere of urban" of an age of "10th" or "20th" in which the gender is "Male".
(14) Setting evidence to "Calm" After setting evidence to "Calm", the result is exhibited in Table 15.We can observe that "Those who have an image of the surrounding area at this shopping street as "Calm" had come with the purpose of visiting for "Shopping" under the image of the surrounding area at this shopping street as "Not fascinating", or "Want to examine deliberately" of an age of "30th", "50th", "60th" or "More than 70" in which the gender is "Female".
(15) Setting evidence to "Atmosphere of urban" After setting evidence to "Atmosphere of urban", the result is exhibited in Table 16.We can observe that "Those who have an image of the surrounding area at this shopping street as "Atmosphere of urban" had come under the image of the surrounding area at this shopping street as "Cheerful", "Want to play" or "Lively" of an age of "10th" or More than 70" in which the gender is "Male".
(16) Setting evidence to "Atmosphere of rural area" After setting evidence to "Atmosphere of rural area", the result is exhibited in Table 17.We can observe that "Those who have an image of the surrounding area at this shopping street as "Atmosphere of rural area" had come with the purpose of visiting for "Business" under the image of the surrounding area at this shopping street as "Not fascinating" or "Want to examine deliberately" of an age of "50th", "60th" or "More than 70" in which the gender is "Female".

REMARKS
The Results for Bayesian Network Analysis are as follows.
In the Bayesian Network Analysis, model was built under the examination of the causal relationship among items.Sensitively Analysis was conducted after that.The main result of sensitively analysis is as follows.We can observe that "Those who have an image of the surrounding area at this shopping street as "Healed" had come under the image of the surrounding area at this shopping street as "Of the united feeling there is", "Cheerful", "Individualistic", "Fascinating", "Want to play", "Lively" or "Atmosphere of urban" of an age of "10th", "20th", "30th" or"40th" in which the gender is "Female".
We can observe that "Those who have an image of the surrounding area at this shopping street as "Fascinating" had come under the image of the surrounding area at this shopping street as "Beautiful", "Of the united feeling there is", "Varied"," Cheerful", "Individualistic", "Friendly", "Healed", "Want to play" or "Lively" of an age of "10th" or "20th" in which the gender is "Female".
We can observe that "Those who have an image of the surrounding area at this shopping street as "Want to play" had come under the image of the surrounding area at this shopping street as "Of the united feeling there is", "Cheerful", "Individualistic", "Friendly", "Healed", "Fascinating", "Lively" or "Atmosphere of urban" of an age of "10th", "20th" or "40th" in which the gender is "Male".
We can observe that "Those who have an image of the surrounding area at this shopping street as "Want to examine deliberately" had come with the purpose of visiting for "Shopping" under the image of the surrounding area at this shopping street as "Ugly", "Gloomy" or "Not fascinating" of an age of "40th", "60th" or "More than 70" in which the gender is "Female".
We can observe that "Those who have an image of the surrounding area at this shopping street as "Lively" had come under the image of the surrounding area at this shopping street as "Beautiful", "Of the united feeling there is", "Varied", "Cheerful", "Individualistic", "Friendly", "Healed", "Fascinating", "Want to play", or "Atmosphere of urban" of an age of "10th" or "20th" in which the gender is "Male".
We can observe that "Those who have an image of the surrounding area at this shopping street as "Atmosphere of rural area" had come with the purpose of visiting for "Business" under the image of the surrounding area at this shopping street as "Not fascinating" or "Want to examine deliberately" of an age of "50th", "60th" or "More than 70" in which the gender is "Female".

Figure 4 .
Figure 4. How often do you come to this shopping street?(Q1)

Figure 5 .
Figure 5.What is the purpose of visiting here?(Q2)

Figure 6 .
Figure 6.How do you feel about the image of the surrounding area at this shopping street?(Q3)

Figure 7 .
Figure 7.There are many old building at the age of nearly 50 years.Do you think we can still use them?(Q4)

Figure
Figure 8.A Built Model

posterior probability APPENDIX 3 Difference of probability
do you feel about the image of the surrounding area at this shopping street?Select the position There are many old building at the age of nearly 50 years.Do you think we can still use them?a. Can use it b.Cannot use it c.Have no idea 5. Is there any functions or facilities that will be useful?