def similarity_metrics(corpus):
model = NGramGraphModel(corpus)
print "Bigram Graph Class"
model.bigram_graph_class()
print "Calculating Similarity Metrics"
corpus = model.calc_sim_metrics()
return corpus
def n_gram_graph_model(data):
#data = pd.DataFrame()
class_labels = ["too localized", "open", "off topic", "not constructive", "not a real question"]
graph_models = {}
#Constroi o modelo em grafo para cada classe
for class_label in class_labels:
data_label = data[data["OpenStatus"] == class_label]
corpus = format_data(data_label)
model = NGramGraphModel(corpus)
graph_model = model.bigram_graph_class()
graph_models[class_label] = graph_model
corpus = format_data(data)
model = NGramGraphModel(corpus)
all_data = model.calc_sim_metrics(graph_models, corpus)
df_data = pd.DataFrame.from_dict(all_data, orient="index")
return df_data