Danger data analysis

Because the technical barriers previously outhned increase uncertainty in the data, plant-performance analysts must approach the data analysis with an unprejudiced eye. Significant technical judgment is required to evaluate each measurement and its uncertainty with respec t to the intended purpose, the model development, and the conclusions. If there is any bias on the analysts part, it is likely that this bias will be built into the subsequent model and parameter estimates. Since engineers rely upon the model to extrapolate from current operation, the bias can be amplified and lead to decisions that are inaccurate, unwarranted, and potentially dangerous. 

Data Analysis Because of the danger of false conclusions if only one or two parameters were evaluated, it was deemed better to correlate every parameter with all the others, and to assemble the results in a triangular matrix, so that trends would become more apparent. The program CORREL described in Section 5.2 retains the sign of the correlation coefficient (positive or negative slope) and combines this with a confidence level (probability p of obtaining such a correlation by chance alone). 

There is a danger in placing too much emphasis on data analysis as such at the expense of data collection and manipulation based on professional experience. A large number of different techniques of analysis are theoretically possible, but often the necessary data are missing. In other words, the tools are available but cannot be used because of the lack of certainty of input data. 

AI research has already provided the concepts of supervised and unsupervised learning to data analysis, and these have proved useful in the classification of analytical methods and to alert us to the potential danger of chance effects. But what of the application of AI techniques themselves... 

European Commission for these and other substances by nominated dates. The toxicological data and estimation of exposure will form the basis of risk analysis and deterrnination of the appropriate restriction and control of substances in the workplace (58). Restriction of the sales of dangerous substances and preparations to the general pubHc is enforced under Directive 76/769 EC (59). 

Suicidal thoughts are one of the symptoms of depression. Paradoxically, one of the best-publicized dangers of SSRIs is their potential to increase the risk of suicide. This heightened risk is especially well established for children, adolescents and young adults. In their most recent analysis of the data, the FDA concluded that, when compared to placebos, SSRIs double the risk of suicidal thoughts and behaviour in depressed patients up to the age of 24/ There also seems to be an increased risk for people who are older than 24, but the interpretation of these data is still disputed.8... 

The comparison of the safety of equipment is not straightforward. It depends on equipment themselves and process conditions. Equipment safety can be evaluated from quantitative accident and failure data and from engineering practice and recommendations. Experience-based information is found from layout recommendations and safety analysis methods such as the Dow E F Index (1987). Quantitative data can be found from accident and operational statistics. It should be remembered that all reliability or failure information does not express safety directly, since all failures are not dangerous and not all accidents are due to failures of equipment. 

The PSI element of both the OSHA PSM Standard and the EPA RMP regulation can be improved by requiring the inclusion of all existing information on chemical reactivity. Examples of such information are chemical reactivity test data, such as DSC, thermogravimetric analysis (TGA), or accelerating rate calorimetry and relevant incident reports from the plant, the corporation, industry, and government. OSHA and EPA should require the facility to consult such resources as Bretherick s Handbook of Reactive Chemical Hazards,Sax s Dangerous Properties of Industrial Materials, and computerized tools (e g., CHETAH, The Chemical Reactivity Work Sheet). 

Regression analysis techniques are then used to find the significant agents and produce a model for the particular data in question. This is essentially an empirical approach and, alternatively, a mathematical form for the data could be found by curve fitting without consideration of the effect of individual agents. Extrapolation is then particularly dangerous. 

---