[SOLVED] algorithm statistic software PRRE5008 Assignment (Total Marks = 100)

PRRE5008 Assignment (Total Marks = 100)
Write a report in which analyses of the following problems are described. This should include a concise summary of the algorithms that were used, the results of the analyses, as well as some conclusions regarding the quality of the solutions. You are free to use software of your choice.
Use the Tennessee Eastman (TE) data and fit a principal component model to the data, using the normal operating condition (NOC) data. More specifically, in each of the worksheets in the Excel data files, use XMEAS_train to fit the model and then XMEAS_test in all the other spreadsheets to test the model.<25>.
Construct and show the Q- and T2-statistics plots for Faults 1, 2 and 3 together with the NOC data.
Construct and show the contribution plots of each of these faults.

Use the TE data and fit a dynamic principal component model to the data, using the same data sets as in (1) <20>.
Again show the Q- and T2-statistics plots for Faults 1, 2 and 3 together with the NOC data.
Also, show the contribution plots of each of these faults.

Fit a classification model to the TE data. Make use of a suitable variant of the Boruta algorithm to identify the variables that could be implicated in each fault condition. In other words, make use of an augmented data sets containing the original variables, as well as a permuted version of each original variable. Repeat this at least twice, so that confidence limits can be estimated for the effect of each of the variables. How does this compare with the results in 1(b)? <30>
Consider the SAKRES data set (SAKRES.xlsx). <25>
Make use of a Shapley regression approach to identify the most important variables contributing to the consumption of the reagent. Also determine which variables do not have a significant impact on the response. A linear model would be fine for this purpose.
Make use of a relative weights regression approach to identify the most important variables contribution to the consumption of the reagent. Also determine which variables do not have a significant impact on the response. How does this compare with the results in 4(a)?
Finally, apply the same approach use in (3) to analyse the contributions of the operational variables to the consumption of the reagent. How does this compare with 4(a) and 4(b)?
Following on from 4(c), provide a partial dependence plot of the effect of each variable that was deemed to have had a significant contribution to the response.
Useful References
Alcala, C.F. and Qin, S.J. 2011. Analysis and generalization of fault diagnosis methods for process monitoring. Journal of Process Control, 21, 322330.
Kourti, T. and MacGregor, J.F. 1995. Process analysis, monitoring and diagnosis, using multivariate projection methods. Chemometrics and Intelligent Laboratory Systems, 28, 3-21.
Kursa, M.B and Rudnicki, W.R. 2010. Feature selection with the Boruta package. Journal of Statistical Software, 36(110, 1-13.
Mishra, M.K. 2016. Shapley value regression and the resolution of multicollinearity. Online at https://mpra.ub.uni-muenchen.de/72116/ MPRA Paper No. 72116, posted 20 Jun 2016 14:09 UTC.
END