Doing Sentiment Analysis using NLP Libraries
Conclusion from the Frist Approach :
Sentiment Analysis using NLP Libraries (Unsupervised learning ) :
result and analysis:
1)AFINN lexicon
Model Performance metrics:
------------------------------
Accuracy: 0.71
Precision: 0.73
Recall: 0.71
F1 Score: 0.71
The Accuracy is 71% and F1 score tell about the performance of the that is 72%.that getting success is 72%.
2)SentiWordnet
Model Performance metrics:
------------------------------
Accuracy: 0.69
Precision: 0.69
Recall: 0.69
F1 Score: 0.68
The Accuracy is 71% and F1 score tell about the performance of the that is 72%.that getting success is 72%.
3)VADER
Model Performance metrics:
------------------------------
Accuracy: 0.71
Precision: 0.72
Recall: 0.71
F1 Score: 0.71
The Accuracy is 71% and F1 score tell about the performance of the that is 72%.that getting success is 72%.
From comparing all three unsupervised model the AFFIN is best model because the precise value is greater than the value VADER .Low value mean the ,no of reparation are less so data varies .the variation is data is large.
Conclusion from the Second Approach :
Sentiment Analysis using NLP Libraries (Supervised learning ) :
result and analysis:
Model Performance metrics:
------------------------------
Accuracy: 0.9
Precision: 0.9
Recall: 0.9
F1 Score: 0.9
The Accuracy is 90% and F1 score tell about the performance of the that is 90%.that getting success is 90%.
Comparing the result of both Learning we find that supervised learning has good result. The Supervised learning accuracy is greater than unsupervised learning model AFFIN .
Comments
Post a Comment