Identifying Relationships Between Eco-innovation and Product Success - Technology Management for Sustainable Production and Logistics - page 180

Table 3 Evaluation of using natural materials for developing new eco-products
Strongly
disagree

Disagree Neither agree

nor disagree

Agree Strongly

agree

Corresponding
semantics

1 2 3 4 5

Respondent 1 0.7 0.3 Partial disapproval

1 Absolutely approval

Respondent 2

0.4 0.6 Partial approval

Respondent 3

Respondent 4 0.8 0.2 Almost total

disapproval

shortcomings of the original scale is that the respondent can make several answers
for each item, and can give each answer as a percentage. If the respondents can
express their judgement as a degree of membership corresponding to the linguistic
variables, it becomes possible to give a real number between 0 and 1. A sample of
answers to the statement “Our unit often emphasizes developing new eco-products
through new technologies to use natural materials” is illustrated in Table 3 .
The weight corresponding to a linguistic variable is different among respond-
ents (employees in an enterprise), since personal preferences are subjective and
fuzzy according to complicated, diverse, and indeterminate human behaviour.

4.2 Artificial Neural Networks for Forecasting

Artificial neural networks (ANNs) are an important class of tool for quantitative
modelling. Today, neural networks are treated as a standard data mining tool and
used for many data mining tasks such as pattern classification, time series analysis,
prediction, and clustering. Neural networks are computing models for information
processing and are particularly useful for identifying the fundamental relationship
among a set of variables or patterns in the data. The popularity of neural networks
is due to their powerful modelling capability for pattern recognition. Several
important characteristics of neural networks make them suitable and valuable for
data mining. First, as opposed to the traditional model based methods, neural net-
works do not require several unrealistic a priori assumptions about the underlying
data generating process and specific model structures. Moreover, the mathematical
property of the neural network in accurately approximating or representing various
complex relationships has been well established. Furthermore, neural networks
are nonlinear models. As real world data or relationships are inherently non-
linear, traditional linear tools may suffer from significant biases in data mining.
Neural networks with their nonlinear and nonparametric nature are more effective
for modelling complex data mining problems. Finally, neural networks are able
to solve problems that have imprecise patterns or data containing incomplete and
noisy information with a large number of variables. This fault tolerance feature is
appealing for data mining problems because real data is usually noisy and does not