Tham khảo tài liệu 'engineering statistics handbook episode 5 part 14', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | . How can I predict the value and and estimate the uncertainty of a single response ENGINEERING STATISTICS HANDBOOK home TOOLS Í AIDS ISEAtCH BACK NEKTl 4. Process Modeling . Use and Interpretation of Process Models . What types of predictions can I make using the model can I predict the value and and estimate the uncertainty of a single response A Different Type of Prediction In addition to estimating the average value of the response variable for a given combination of preditor values as discussed on the previous page it is also possible to make predictions of the values of new measurements or observations from a process. Unlike the true average response a new measurement is often actually observable in the future. However there are a variety of different situations in which a prediction of a measurement value may be more desirable than actually making an observation from the process. Example For example suppose that a concrete supplier needs to supply concrete of a specified measured strength for a particular contract but knows that strength varies systematically with the ambient temperature when the concrete is poured. In order to be sure that the concrete will meet the specification prior to pouring samples from the batch of raw materials can be mixed poured and measured in advance and the relationship between temperature and strength can be modeled. Then predictions of the strength across the range of possible field temperatures can be used to ensure the product is likely to meet the specification. Later after the concrete is poured and the temperature is recorded the accuracy of the prediction can be verified. V fix. 3i The mechanics of predicting a new measurement value associated with a combination of predictor variable values are similar to the steps used in the estimation of the average response value. In fact the actual estimate of the new measured value is obtained by evaluating the estimated regression function at the .
