Differences

This shows you the differences between two versions of the page.

--- forecast_tool_captrd [2020/02/27 11:56] – [Step 4: Breaking down results from Member State to regional and farm type level] matsz
+++ forecast_tool_captrd [2022/11/07 10:23] (current) – external edit 127.0.0.1
@@ Line 477: / Line 477: @@
 \begin{equation}
-X_{MS,i,t}^{GROF,Trend}=\sum_{r\inMS}X_{r,i,t}^{GROF,Trend}
+X_{MS,i,t}^{GROF,Trend}=\sum_{r\in MS}X_{r,i,t}^{GROF,Trend}
 \end{equation}
 \begin{equation}
-X_{MS,"levl",t}^{j,Trend}=\sum_{r\inMS}X_{r,"levl",t}^{j,Trend}
+X_{MS,"levl",t}^{j,Trend}=\sum_{r\in MS}X_{r,"levl",t}^{j,Trend}
 \end{equation}
 \begin{equation}
-X_{MS,"levl",t}^{j,Trend}\cdot \left(X_{MS,"feed",t}^{j,Trend}+10 \right)=\sum_{r\inMS}X_{r,"levl",t}^{j,Trend}\cdot \left(X_{MS,"feed",t}^{j,Trend}+10 \right)
+X_{MS,"levl",t}^{j,Trend}\cdot \left(X_{MS,"feed",t}^{j,Trend}+10 \right)=\sum_{r\in MS}X_{r,"levl",t}^{j,Trend}\cdot \left(X_{r,"feed",t}^{j,Trend}+10 \right)
 \end{equation}
@@ Line 501: / Line 501: @@
 **Table 24: Example of the final output of the trends estimation task and description of the variables**
+^Product code^  Activity code  ^  Variables  ^  Years  ^^^^^^^^^  Explanations  ^
+^  ^ ^ ^1984^…^2009^2010^2011^2012^2013^2014^2015^:::^
+^SWHE^	SWHE^	BASM	| 	 | | | | | | |	|   	 	 	 	 	8337|Base year value from Build database workstep.|
+^ ^ ^Penalty	| 	 | | | | | | |	|  	 	 	 	 	 	0.2|"squared root" difference between actual estimate and support value. The larger the value, the farer the estimate from support.|
+^ ^ ^Lo	 	 | 	 | | | | | | |	|  	 	 	 	 	 	8080|	Lower estimation bound.|
+^ ^ ^	 	DGAgri1	 | | |	 	8876|	8385	|8046	|8109|	8632|	8996|	9167|	Projection of Aglink-Cosimo for the EU15 aggregate scaled to fit the CAPRI database.((Aglink-Cosimo model produces projections not for each EU MS, but for the EU aggregates: EU, EU "old" MSs and EU "new" MSs. During the calibration process these values are first scaled to better fit the CAPRI database. These scaled values are then used for the calibration procedure.))|
+^ ^ ^ 	 	TrustLevl	| 	 | | | | | | |	|  	 	 	 	3|	Exogeneous value used for restricting min and max values of the support values. It is used in calculating lower and upper bounds (up and lo) of the projections.|
+^ ^ ^ 	 	data	 | 	 | | | | | | |	| | |
+^ ^ ^	 	BAST	 | 	 | | | | | | |	|  	 	 	 	8579|	Simple average of the last 3 observation years available: 2012-2014.|
+^ ^ ^ 	 	B2000	| 	 | | | | | | |	|  	  	 	 	7988| |
+^ ^ ^ 	 	support	 | 	 | | | | | | |	|  	 	 	 	9167|	Values estimated as linear combination of Step1 and BAST (BASM) with R2 as weight. They are replaced with expert support where applicable and then scaled. They are then stored as Support1. Support is then redefined based on the Aglink-Cosimo value.((The final version of the support value at MS level (if calibration to the projections of Aglink-Cosimo takes place), is calibration value derived from DgAgri1.))|
+^ ^ ^	 	support1	| 	 | | | | | | |	|    	 	 	8943|	(expert) support value, before introduction of Aglink-Cosimo calibration values. |
+^ ^ ^ 	 	step1	 	| 	 | | | | | | |	|  	 	 	8918|	1) Result of estimation of unconstrined trends|
+^ ^ ^ 	 	step2	 	| 	 | | | | | | |	|  	 	 	8851|2) Results of solving the trend model with constraints at MS level and with support1|
+^ ^ ^ 	 	step3	 | 	 | | | | | | |	|  	 	 	 	8949|3) First, it is defined as results of solving trend model with constraints at MS level and with support (defined with Aglink-Cosimo value). Then, it is redifined with the results from solving this trend model with additional constraints at NUTS2 level.  |
+^ ^ ^ 	 	wVarErr	 | 	 | | | | | | |	|  	 	 	 	259353|	Error variance.
+^ ^ ^	 	CoefVarErr	| 	 | | | | | | |	| 	 	 	0.1| |
+^ ^ ^	 	Extrap	 	 | 	 | | | | | | |	| |  |
+^ ^ ^ 	 	Longrun| 	 | | | | | | 	8553 | 8579	| 	8633| |
+^ ^ ^ 	 	Longrun1	| 	 | | | | | | |	| | |
+^ ^ ^ 	 	P_Data	|6975|	…|	9061|	8614|	8078	|8139|	8810|	8789|  |	Historical data – output of Build database. The last observation year – 2014.|
+^ ^ ^ 	 	series	|6975|	…	|9061	|8614	|8078	|8139|8810	|8789	|  8949| 	Historical values (until 2014) and projected values (starting from 2015 with 5-year step, as defined in the GUI setting for the Trends projection task). The projected values are "copied" from Step 3.((Because the last observation year is 2014, values in 2015 are prjections.))((Values in 2020, 2025 and 2030 are projections as well but are not presented in this example.))|
+^ ^ ^ 	 	up	| 	 | | | | | | |	|     	 	 	 	 	8978	|Upper estimation bound|
+Source: own compilation. Comments: SWHE in Product code column indicates soft wheat commodity. SWHE in Activity code indicates yield of soft wheat. The CAPRI model used for this example was calibrated to the projections of Aglink-Cosimo model.