Per #1749, update the documentation about the HSS_EC statistic.

dtcenter · May 28, 2021 · 98d1f17 · 98d1f17
1 parent 159b393
commit 98d1f17
Show file tree

Hide file tree

Showing 4 changed files with 40 additions and 1 deletion.
diff --git a/met/docs/Users_Guide/appendixC.rst b/met/docs/Users_Guide/appendixC.rst
@@ -240,7 +240,7 @@ HK is also known as the True Skill Statistic (TSS) and less commonly (although p
 Heidke Skill Score (HSS)
 ~~~~~~~~~~~~~~~~~~~~~~~~
 
-Called "HSS" in CTS output :numref:`table_PS_format_info_CTS`
+Called "HSS" in CTS output :numref:`table_PS_format_info_CTS` and "HSS" in MCTS output :numref:`table_PS_format_info_MCTS`
 
 HSS is a skill score based on Accuracy, where the Accuracy is corrected by the number of correct forecasts that would be expected by chance. In particular, 
 
@@ -252,6 +252,21 @@ where
 
 HSS can range from minus infinity to 1. A perfect forecast would have HSS = 1.
 
+Heidke Skill Score - Expected Correct (HSS_EC)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Called "HSS_EC" in MCTS output :numref:`table_PS_format_info_MCTS`
+
+HSS_EC is a skill score based on Accuracy, where the Accuracy is corrected by the number of correct forecasts that would be expected by chance. In particular,
+
+.. math:: \text{HSS } = \frac{n_{11} + n_{00} - C_2}{T - C_2},
+
+where
+
+The C_2 value is user-configurable with a default value of T divided by the number of contingency table categories.
+
+HSS_EC can range from minus infinity to 1. A perfect forecast would have HSS_EC = 1.
+
 Odds Ratio (OR)
 ~~~~~~~~~~~~~~~
 

diff --git a/met/docs/Users_Guide/config_options.rst b/met/docs/Users_Guide/config_options.rst
@@ -1966,6 +1966,22 @@ intensive and slows down the runtime significantly.
 		
   rank_corr_flag = FALSE;
 
+**hss_ec_value**
+
+The "hss_ec_value" entry is a floating point number used in the computation
+of the HSS_EC statistic in the MCTS line type. It specifies the expected
+correct (EC) rate by chance for multi-category contingency tables. If set
+to its default value of NA, it will automatically be replaced with 1.0
+divided by the MCTC table dimension. For example, for a 4x4 table, the
+default hss_ec_value is 1.0 / 4 = 0.25.
+
+It set, it must greater than or equal to 0.0 and less than 1.0. A value of
+0.0 produces an HSS_EC statistic equal to the Accuracy statistic.
+
+.. code-block:: none
+		
+  hss_ec_value = NA;
+
 **duplicate_flag**
 
 The "duplicate_flag" entry specifies how to handle duplicate point

diff --git a/met/docs/Users_Guide/grid-stat.rst b/met/docs/Users_Guide/grid-stat.rst
@@ -225,6 +225,7 @@ __________________________
   mpr_thresh     = [];
   eclv_points    = 0.05;
   rank_corr_flag = TRUE;
+  hss_ec_value   = NA;
   tmp_dir        = "/tmp";
   output_prefix  = "";
   version        = "VN.N";

diff --git a/met/docs/Users_Guide/point-stat.rst b/met/docs/Users_Guide/point-stat.rst
@@ -338,6 +338,7 @@ ________________________
   mpr_thresh     = [];
   eclv_points    = 0.05;
   rank_corr_flag = TRUE;
+  hss_ec_value   = NA;
   sid_inc        = [];
   sid_exc        = [];
   duplicate_flag = NONE;
@@ -914,6 +915,12 @@ The first set of header columns are common to all of the output files generated
   * - 38-40
     - GER, :raw-html:`<br />` GER_BCL, :raw-html:`<br />` GER_BCU
     - Gerrity Score and bootstrap confidence limits
+  * - 41-43
+    - HSS_EC, :raw-html:`<br />` HSS_EC_BCL, :raw-html:`<br />` HSS_EC_BCU
+    - Heidke Skill Score with user-specific expected correct and bootstrap confidence limits
+  * - 44
+    - EC_VALUE
+    - Expected correct rate for HSS_EC
 
 .. _table_PS_format_info_PCT: