Merge pull request #4 from ottenbreit-data-science/bugfix

mathias-von-ottenbreit · web-flow · commit 79fb9bd455bb · 2022-06-15T20:56:28.000+02:00
Bugfix in bin splitting
diff --git a/cpp/term.h b/cpp/term.h
@@ -367,14 +367,16 @@ void Term::setup_bins()
         bins_start_index.push_back(0);
         if(bins>1)
         {
-            for (size_t i = min_observations_in_split-1; i <= max_index+1-min_observations_in_split; ++i) //for each observation in an allowable range
+            size_t start_row{min_observations_in_split-1};
+            size_t end_row{max_index+1-min_observations_in_split};
+            for (size_t i = start_row; i <= end_row; ++i)
             {
                 size_t last_bin_start_index{bins_start_index[bins_start_index.size()-1]};
-                bool eligible_on_spacing_between_observations{i >= last_bin_start_index + observations_in_bin || values_sorted_unique.size()<=bins};
+                bool eligible_on_spacing_between_observations{i >= last_bin_start_index + observations_in_bin || values_sorted_unique.size()<=bins || i == start_row || i == end_row};
                 bool eligible_on_unique_numbers{i>0 && !check_if_approximately_equal(sorted_vectors.values_sorted[i],sorted_vectors.values_sorted[i-1])};
 
                 bool create_bin{eligible_on_spacing_between_observations && eligible_on_unique_numbers};
-                if(create_bin) 
+                if(create_bin)
                 {
                     bins_start_index.push_back(i);
                 }
diff --git a/setup.py b/setup.py
@@ -15,7 +15,7 @@
 
 setuptools.setup(
     name='aplr',
-    version='1.0.4',
+    version='1.0.5',
     description='Automatic Piecewise Linear Regression',
     ext_modules=[sfc_module],
     author="Mathias von Ottenbreit",

Original file line number	Diff line number	Diff line change
`@@ -367,14 +367,16 @@ void Term::setup_bins()`
`367`	`367`	`bins_start_index.push_back(0);`
`368`	`368`	`if(bins>1)`
`369`	`369`	`{`
`370`		`- for (size_t i = min_observations_in_split-1; i <= max_index+1-min_observations_in_split; ++i) //for each observation in an allowable range`
	`370`	`+ size_t start_row{min_observations_in_split-1};`
	`371`	`+ size_t end_row{max_index+1-min_observations_in_split};`
	`372`	`+ for (size_t i = start_row; i <= end_row; ++i)`
`371`	`373`	`{`
`372`	`374`	`size_t last_bin_start_index{bins_start_index[bins_start_index.size()-1]};`
`373`		`- bool eligible_on_spacing_between_observations{i >= last_bin_start_index + observations_in_bin \|\| values_sorted_unique.size()<=bins};`
	`375`	`+ bool eligible_on_spacing_between_observations{i >= last_bin_start_index + observations_in_bin \|\| values_sorted_unique.size()<=bins \|\| i == start_row \|\| i == end_row};`
`374`	`376`	`bool eligible_on_unique_numbers{i>0 && !check_if_approximately_equal(sorted_vectors.values_sorted[i],sorted_vectors.values_sorted[i-1])};`
`375`	`377`
`376`	`378`	`bool create_bin{eligible_on_spacing_between_observations && eligible_on_unique_numbers};`
`377`		`- if(create_bin)`
	`379`	`+ if(create_bin)`
`378`	`380`	`{`
`379`	`381`	`bins_start_index.push_back(i);`
`380`	`382`	`}`