Update model based on a few false positives

README.md

@@ -135,12 +135,18 @@ export ELEIA_Logging__LogLevel__Default=Error
 It is using ML.NET framework to perform machine learning-based binary classification,
 and AutoML (`mlnet auto-train`) was used to generate model.
 
+Exact run:
+
+```batch
+mlnet auto-train -T binary-classification -d trainingdata2.tsv -o ML --label-column-name code -x 180
+```
+
 [trainingdata2.tsv](https://github.com/ktos/Eleia/blob/master/trainingdata2.tsv)
 was a file which the training was performed on.
 
 [trainingdata2.log](https://github.com/ktos/Eleia/blob/master/trainingdata2.log)
 is a log of auto-training, SdcaLogisticRegressionBinary was decided with
-accuracy of 0.9585.
+accuracy of 0.9572.
 
 I was also running longer training sessions than 180 seconds, but still that
 algorithm was decided as the best, longer time had no visible impact.

README.pl.md

@@ -138,11 +138,17 @@ export ELEIA_Logging__LogLevel__Default=Error
 Został wykonany ML.NET do wykorzystania modelu binarnej klasyfikacji, natomiast
 AutoML (tj. `mlnet auto-train`) zostało wykorzystane do wygenerowania modelu.
 
+Konkretnie:
+
+```batch
+mlnet auto-train -T binary-classification -d trainingdata2.tsv -o ML --label-column-name code -x 180
+```
+
 [trainingdata2.tsv](https://github.com/ktos/Eleia/blob/master/trainingdata2.tsv)
 to plik, na którym model był trenowany.
 
 [trainingdata2.log](https://github.com/ktos/Eleia/blob/master/trainingdata2.log)
-to log z autotreningu, SdcaLogisticRegressionBinary z dokładnością 0,9585 został
+to log z autotreningu, SdcaLogisticRegressionBinary z dokładnością 0,9572 został
 ostatecznie wykorzystany.
 
 Uruchamiałem też dłuższe sesje trenujące niż 180 sekund, ale nadal ten algorytm

test/PostAnalyzerTests.cs

@@ -96,11 +96,11 @@ public void Analyze_PostWithLinksThreshold08_ProblemsFound()
         }
 
         [Fact]
-        public void Analyze_PostWithUnformattedCodeThreshold098_UnformattedCodeFound()
+        public void Analyze_PostWithUnformattedCodeThreshold097_UnformattedCodeFound()
         {
             var myConfiguration = new Dictionary<string, string>
             {
-                {"threshold", "0.98"}
+                {"threshold", "0.97"}
             };
 
             var configuration = new ConfigurationBuilder()