302a3 - you will have to do a PCA for this. Notice that the...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
2009 Assignment 3. Hand in by 1.00pm Monday 18 th May. Use the same data set as in Assignment 1. Remove the influence of tot_pop from tot_boats, tot_fisher, and FF_catch. i.e. convert them to numbers per member of population. Drop tot_pop from the data set. Use the log transformation as in the earlier assignments. Then scale (standardise) the transformed variables to unit standard deviation, because we do not want variables with larger variance (on the log scale) to dominate the results) – they are very different variables. Perform a PCO on a Manhattan distance matrix on the transformed data set. . 1) Plot a scree diagram. Comment on what you see. 2) Plot the scores on the first 2 principal axes (the reduced space plot) using the potential impact of fishing (the group variable) as the plotting symbol. Describe any patterns you can see. 3) Now do the same for the first two principal components of the same data. Yes,
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: you will have to do a PCA for this. Notice that the data have been standardised to unit SD, so this is doing the analysis on a correlation matrix. Compare the two plots. Comment on what you see. 4) What proportion of the total variation is preserved by the PCO reduced space plot? Use an appropriate measure. 5) Present the PA correlations for the first 2 PAs. Interpret them. Compare them with the PC component correlations. Comment on the comparison. 6) Plot a couple of internal variables as bubble plots. Interpret what you see. 7) Perform a non-metric multidimensional scaling, for a number of dimensions and plot a STRESS diagram. How adequate is the 2 dimensional plot? What might be a more appropriate number of dimensions? 8) Plot the group variable into the reduced space plot from the non-metric MDS. Compare with the same plots from the PCO and the PCA. Comment on what you see....
View Full Document

This note was uploaded on 09/24/2009 for the course STATS 731 taught by Professor Renate during the Spring '09 term at Auckland.

Ask a homework question - tutors are online