Véronis Pre-Sets, seem to work best (smaller corpus also had better results)

src/config.py

@@ -8,6 +8,7 @@ Choose paths for corpus, dataset and output.
 '''
 corpus = "/proj/absinth/wikipedia_reduced/"
 dataset = "../WSI-Evaluator/datasets/MORESQUE/"
+test = "../WSI-Evaluator/datasets/trial/"
 output = "../output/"
 
 '''
@@ -15,14 +16,14 @@ Choose stop words and allowed pos-tags.
 - Stop words will not be considered for nodes.
 - Only tokens with allowed pos-tags will be considered.
 '''
-stop_words = ['utc', "'s", 'new', 'other', 'talk', 'wikipedia', 'article', 'topic', 'page', 'editors', 'encyclopedia', 'free']
-allowed_tags=['NN','NNS','JJ','JJS','JJR','NNP','VBZ','VBG']
+stop_words = ['utc', "'s", 'new', 'other', 'talk', 'wikipedia', 'article', 'topic', 'page', 'editors', 'encyclopedia', 'free', 'pp']
+allowed_tags = ['NN','NNS','JJ','JJS','JJR','NNP']
 
 '''
 Choose the maximum number of nodes and edges that should be considered before building the graph.
 '''
-max_nodes = 100000
-max_edges = 10000000
+max_nodes = 20000
+max_edges = 2000000
 
 '''
 Choose the minimum context size.
@@ -43,4 +44,4 @@ Choose minimum number of neighbors and maximum median weight of the most frequen
 - the threshold is calculated using the media of the same number of neighbors declared in min_neighbors.
 '''
 min_neighbors = 6
-theshold = 0.8
+threshold = 0.8