Text-based emotion recognition in online social networks using adaboost classification method

Paria Soheilifar, Samad Nejatian, Karmollah Bagherifard

Abstract


Data mining and natural language processing are used in Emotion Mining to retrieve and extract knowledge from text. Data is always changing as a result of upgrades and the inclusion of new information at any given time. The feature set can be reduced to a feature selection method, which can be reduced to a subset with considerably smaller volume and higher detection capabilities, using the new feature combination. As you may be aware, synonymous words are treated differently in text classification. The composition method's primary assumption is that combining synonyms results in better characteristics. Given the complexity of the search problem, the use of meta-meta-methodologies can be beneficial in identifying better combinations of features and, as a result, enhancing classification efficiency. The multipurpose method of learning and optimization (TLBO) algorithm will be employed for this project because to its simplicity and quickness in identifying individuals' perspectives. In the field of cognitive science, the proposed strategy has a considerable effect on data reduction and, as a result, classification efficiency. In this study, we used Adaboos, a hybrid classification approach, to classify comments. Adequacy should improve classification accuracy by an average of 8 and 11 points over regression and SVM, respectively.

Keywords


Adaboost; Emotion Mining; Feuture Union; Hyprid Classification; TLBO



DOI: http://doi.org/10.11591/ijict.v12i2.pp%25p

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

View IJICT Stats