Page 331 - Solid Waste Analysis and Minimization a Systems Approach

PERFORMANCE PARAMETER CASE STUDY 309

As shown in the validation table, all the regression coefficients developed from the
research data (national survey) and data gathered from other agencies were statistically
equal at the 95 percent confidence level. This validated the research data and model
findings with an external source.
The integrated model was also validated using an artificial intelligence (AI) pro-
gram (neural networks) to examine one typical waste group with a relatively large
amount of research data. The commercial/government waste group was chosen
with 124 company waste records. The results of the AI program yielded similar
results to the multivariable regression model developed for this waste group (an
average error estimate of 5.8 percent). The benefits of AI are the ability for the pro-
gram to learn as new data is entered to strengthen the predictions. One drawback
of AI for this research is that larger amounts of data are required over that of regres-
sion modeling.

19.3 Demonstration of the Prediction

of Solid Waste Generation Using the

Developed Model

This section discusses a demonstration case study of the integrated environmental
model’s prediction capability and margin of error. An electronic manufacturing equip-
ment company was randomly selected from the Greene County data set. Table 19.3
displays the actual data that was gathered from the company and the values predicted
by the integrated environmental model. The company employed 60 people, paid $60
per ton to dispose of solid waste, and was not ISO 14001 certified.
As shown in the table, the actual data was 13.3 tons or 5.4 percent more than the
model prediction. At most, actual material composition tonnages were approxi-
mately 10 percent different than predictions with most approximately +/− 5 percent
different. This random sample demonstration indicated the effectiveness and accu-
racy of the integrated environmental model. Figure 19.1 displays the graphical version
of the data.

19.4 Performance Parameter

Case Study

Two companies were selected from the electronic manufacturer’s waste group to
demonstrate the solid waste performance parameters. These companies’ waste records
were provided by the Greene County (Ohio) Solid Waste District. One company