On my TODO list was to start learning R, including how to do R from within Perl. So I got Statistics::R installed.
Not really sure of what kind of model to apply to any of this data, what I decided to try was Hyndman's ets() from the forecast package at CRAN. This is a swiss army chainsaw of exponential smoothing models. The first (only so far) set of data I through at it, was the CPU Load (%). As mentioned earlier, the median CPU load was 100%. The ets() function will choose the best from a large assortment of models. But by and large what I've seen is that it always went to the same model (no linear trend, no seasonality), with a smoothing factor so small that fitted values only do a small random walk around the mean CPU Load. Forcing ets() to use a significantly larger smoothing factor, does produce larger swings in this random walk about the mean, but the sum of residuals squared does actually seem to get worse. The sets of data which are in Watts (instead of CPU Load %) should behave the same way as the CPU Load, so no sense looking at them. The temperature data should be different. This shouldn't show a trend, but it could show seasonality. What I would expect is that the smoothed data temperature data should show a correlation with the smoothed power, except with a lag. Gord _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
