Page 139 - Contributed Paper Session (CPS) - Volume 2
P. 139
CPS1474 Jing R. et al.
the scientific nature of the collection method, computer-aided telephone
survey (CATI) is adopted.
The survey questions about the above six aspects, and asked consumers
about their current situation and expectations for the next three months for
each. Our questionnaire is located in the form of a five-level scale, that is, each
question has five responses: extremely positive, more positive, neutral,
negative, and extremely negative. The actual results of surveys over the years
show that the reliability of each dimension in the customer’s confidential
questionnaire is higher than 0.8, indicating that the questionnaire has a high
reliability. The absolute estimation error also can be controlled within ±1.3 on
the condition of 95% confidence level.
By summarizing the survey data covering six aspects, the total customer’s
confidence index and each sub-item eventually obtained. Where, the total
index is the weighted average of satisfied index and expected index:
CCI = Satisfied index ×40%+ Expected index×60%
The weights are also applicable for sub-items of economy, employment,
price, living, real assets and investment. Total satisfied index is obtained by
evaluating mean of six sub-satisfied index, the same is true for expected index.
Each sub-satisfied index is obtained by the follow equation:
Sub-satisfied (expected) index = 100 + [ (100×extremely positive%) +
(50×more positive%) – (50×negative%) + (100×extremely negative%)]
It is thus obvious that, the value of CCCI ranges from 0-200, 0 ≤CCI≤200,
with 0 being no confidence and 200 being most confident. And CCI = 100 is
the demarcation line of customer’s confidence, more than 100 shows
confidence is positive, otherwise is negative. The larger the value, the stronger
confidence of consumers, and the weaker the opposite.
State Space Model
The existing research analysed the trends of CCCI and its relationships with
each sub-item by descriptive statistics [2,8]. However, cause of repeated
information between each sub-item, their respective status and roles in the
CCCI are different [9]. So, this paper inspects relationships between CCCI and
each sub-item, to demonstrating the evolution of CCCI and its causes in the
past decade. As time progresses, these relationships are constantly changing,
it is impossible to simply calculate the "on average" influence of sub-items on
CCCI. Therefore, we introduce a State Space Model (SSM) with variable
parameter to reflect dynamic relationships between them.
SSM consists of a set of state equations and observation equations. The
main advantages are: SSM can integrate observable and unobservable
variables in observation equations, and then estimating them together; the
128 | I S I W S C 2 0 1 9