Univariate Model Forecasting was used to develop a forecast of transaction arrival rates by performance groups/periods that were identified in the case study discussed in Section 3.3.3.2. Refer to Figure 3-56 to consider the historical observations for Workload 01 (denoted by WKL01). There is a significant decrease in the transaction arrival rate for the workload in the week of September 11th. The week contains a one-day holiday, thus you can reasonably expect a decrease in the volume of the workloads for the week. Therefore, a linear model was proposed for the workload and deleted the observation for the week of September 11th. Figure 7-19 shows the control parameters that were supplied to produce the model.
/------------------------ Univariate Model Forecasting -----------------------\ |Command ===> | | | |Composing CA MICS Inquiry: PRFUVM - Univariate Modeling Inquiry | | | |Report title ===> Univariate Modeling Inquiry | | | |Selection criteria: | | Input file ===> PRF - Performance Group Example | | Start date ===> _________ (ddmonyyyy) | | SYSID ===> TSO1 | | Zone ===> 1________ | | No. months ===> ____ (1-9999) | | Save forecast ===> YES (YES/NO/AGE) | | Specify CAPAPU Values ===> N (Y/N) | |Forecast data element ===> WKL01 | |Forecast length (months) ===> 4 (1-9999) | |Confidence limits (percent) ===> 95 (70/90/95) | | | |Time element terms: Linear ===> YES (YES/NO) | | Quadratic ===> ___ (YES/NO) | | Cubic ===> ___ (YES/NO) | |Specify extended options ===> Y (Y/N) | \------------------------------------------------------------------------------/ Figure 7-19. Univariate Model Forecasting Control Parameters (Screen 1 of 2) /-------------------- Delete Historical Data Observations --------------------\ |Command ===> Scroll ===> CSR | | | |Inquiry Step: Univariate Model Forecasting | | | |Line Cmds: I Insert D Delete R Repeat M Move C Copy | | | |Cmd --------- Dates to be dropped (specify in ddmonyy format) ---------- | | - ------- ------- ------- ------- ------- ------- ------- ------- | | _ 11SEP98 _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | | _ _______ _______ _______ _______ _______ _______ _______ _______ | \------------------------------------------------------------------------------/ Figure 7-19. Univariate Model Forecasting Control Parameters (Screen 2 of 2)
Figure 7-20 shows the Residual Analysis Report for the model. As you can see, all of the predicted values are within 5 to 7% of the actual values except for the observation recorded for the 11th of September. Since this week was not included in the development of the model, the residual value for this week does not exist.
CA MICS Capacity Planner UNIVARIATE MODELING RESIDUAL ANALYSIS UNIVARIATE FORECAST OF: WKL01 BASED ON TIME ELEMENTS: LINEAR TRANSFORMED PREDICTED RESIDUAL -------------PLOT OF RESIDUALS------------- CONFIDENCE DATE WKL01 WKL01 WKL01 WKL01 -2334 0.0 +2334 LIMITS +/- ------- ---------- ------------ ---------- ---------- ------------------------------------------- ---------- 24JUL98 11102.50 11102.50 10688.77 413.73 | 0+++ | 986.87 31JUL98 11026.70 11026.70 10763.89 262.81 | 0++ | 812.25 07AUG98 11232.20 11232.20 10839.02 393.18 | 0+++ | 667.92 14AUG98 10096.80 10096.80 10914.14 -817.34 | -------0 | 577.08 21AUG98 10604.80 10604.80 10989.26 -384.46 | ---0 | 566.08 28AUG98 10192.30 10192.30 11064.38 -872.08 | -------0 | 639.08 04SEP98 11505.60 11505.60 11139.50 366.10 | 0+++ | 772.61 11SEP98 8881.00 . 11214.62 . | 0 | 941.25 18SEP98 11927.80 11927.80 11289.74 638.06 | 0+++++ | 1129.39 26SEP98 . . 11364.86 . | 0 | 1328.77 03OCT98 . . 11439.98 . | 0 | 1535.01 10OCT98 . . 11515.11 . | 0 | 1745.69 17OCT98 . . 11590.23 . | 0 | 1959.37
Figure 7-20. Univariate Forecast of Performance Group/Period Workloads
|
Copyright © 2014 CA.
All rights reserved.
|
|