This is a try to evaluate a living environment synthesized using the discrimination method with the theory of quantification II which is multi valuable factor analysis.
Classes of the outsider were “good”, “neutral” or “bad”. The scores were obtained for categories of factors which affect the synthesized evaluation of a living environment with the theory of quantification II. The dividing points for discrimination were obtained on the mini-max method. They were used to predict a class of the outsider with a total score.
Here it is omitted to tell every detail of the social survey, but we distributed questionnaires with the effort to find places which result in even distribution on each category for each factor. As a result, the period of survey was for 11 years to have 540 answers.
Table 1 Category scores
for combined rating of a living environment
At the questionnaires we got 23 factors for
a living environment but most
of present residential areas in Japan are in city types and the internal
correlation between each factor in twenty-three was high. They were reduced to
eight factors with the least internal correlation to be applicable for the
condition of the theory. Namely, each factor must be independent.
The obtained result on the theory is given in Table 1. Categories of a factor were collected to have less correlation ratio in Table 1.
How well the outsider is explained with the factors used is given by a correlation ratio.
It has the dimension of variance and its square root is obtained for the dimension of a correlation coefficient. It was 0.73.
How deeply each factor is correlated to the outsider is given by a partial correlation coefficient. Each factor’s partial correlation coefficient is shown in Fig.1.
Fig.1 Partial correlation coefficient of each factor
When the outsider has three classes “good”, “neutral” or “bad”, it needs points where it should be distinguished on the total score. They should be given to have the highest hit provability. Here the mini max method was applied when the provability to fall on each class is unknown. Each area of distribution was divided to be equal between two neighboring dividing points there the frequency distribution was imagined to have the normal distribution.
Thus obtained dividing points are shown in Fig. 2, the one between "bad" and "neutral” was Z1=-0.38 and the one between “neutral” and “good” was Z2=1.03. The provability calculated by the area was 68% and the one obtained from the original data was 69%.
The normal distribution of each class and the frequency distribution of the original data are shown with the distribution points.
Fig.2 Dividing points for combined rating for a living environment
If this method is applied to the experimental house at Kaiwaka,
Convenience for shopping-bad;-0.251
Sunny in winter-very good; 0.303
Light in the night-bad; -0.113
Noise-good; 0.241
Safety for commute and school-good; 0.023
Greens around-very good; 0.216
Public moral around-good; 0.382
Safety for children’s play-neutral; -0.068
, and the total score is 0.733. It is referred to the dividing points in Figure 2 and it is predicted “neutral”. It is very close to “good”. For instance, if the convenience for shopping is improved, the living circumstances are estimated as “good”. The result coincides to my daily impression.
Living environment along a high way with high noise level
We got an opportunity to survey living environments exposed with high noise level along the high way in Osaka (from the 8th July to the 3rd August in 1983). The median value of the noise level was 65 dB(A) during the night. About 500 answers were sent back with 90% of collection.
Fig.3 shows their responses to their living environments.
From the total score of each answer, a class of the outsider was predicted by the method given in the previous section and is shown in the left. The real responses are given in the square. For instance, 84 were predicted to be good but responded answers got 50 for “good” and 34 for “neutral”
The prediction coincided with 65%. The hitting provability given in the previous section was 68% and this method shows practically usable.
Fig.3 Evaluation of their living environments in the left and their responded results in the square.
Towards the improvement of the living environments in this area
If noise environment is very bad, it takes -0.578 from the score in Fig.1and it needs 0.715(=0.137+0.578) to improve the noise environment to neutral (0.137).
When a factor of Shopping convenience, Night lighting, or Greens has neutral, either of them can not improve the high noise environment by itself even it gets very good environment.
The importance of the noise control at the source is clearer as it has been suggested after it is compared with other living environments.
When noise environment is improved to neutral and the gained score is added to the answered results, synthesized evaluation along the high way changed to higher classes. How many of them were changed is shown in Fig.4.
Fig.4 How many have changes to higher classes after the noise environment is changed to neutral.
Every class gets higher one by the improvement of noise environment. Even other living environments are kept same, 27% of people get “good” and 83% of people get “neutral”
It shows that they need really the improvement of the noise environment.
Proposal of a joint tunnel
A joint tunnel can have electrical cables, civic gas pipes, telephone lines, information lines of optic fiber, sewage system, water lines, waste conveyance and so on. Suppose a highway is berried underground with a joint tunnel having the green belt on top.
The area along the high way gets 0.715(=0.137+0.578) from the noise environment improvement and 0.929(=0.708+0.212) from greens and is expected to get 1.644 in total.
If they are in bad living environment to have －0.800, they get 0.844 in the total score having the joint tunnel. If the score is referred to Fig.2, the living environment becomes better to “neutral” besides it is very close to “good”.
They have a certain level of satisfaction on other factors in such a city type area. It is natural that they get higher evaluation by the improvement of noise and greens environments.
Thus, I propose the introduction of the joint tunnel.