Figure 23From: Enriching feature engineering for short text samples by language time series analysisPosterior distributions and 95% credible intervals of differences for log loss scores from different methods. (a) Time Series Method Alone vs NLP NSC Method Alone. (b) NLP MNB Method Alone vs Time Series Method Alone. (c) NLP MNB Method Alone vs NLP NSC Method Alone. (d) NLP MNB Combined with Time Series Features vs NLP MNB Method Alone. (e) NLP NSC Combined with Time Series Features vs NLP NSC Method AloneBack to article page