TY - GEN
T1 - Intent classification of social media texts with machine learning for customer service improvement
AU - Pérez-Vera, Sebastián
AU - Alfaro, Rodrigo
AU - Allende-Cid, Héctor
N1 - Publisher Copyright:
© Springer International Publishing AG 2017.
PY - 2017
Y1 - 2017
N2 - Social media platforms in the last few years have facilitated the development of communities that discuss real-world events, and have shaped the way users interact. The content generated in these platforms reflect a variety of intentions, ranging from social interaction to commercial interest, among many others. The present study aims at the implementation of an automatic intent classification system for a Chilean electricity company social media account. The dataset was created from 5000 tweets that were manually classified by 5 people. If discrepancies were detected, a majority voting scheme was used in order to tag the tweets’ intentions. In order to perform the experimental validation of the automatic classification with the machine learning algorithms, several text representations were used (tf-idf, tf-rfl and bin-rfl). The results obtained from the various tests that were conducted yielded satisfactory results. We also analyzed how to assign automatic responses to frequently asked questions, and obtained promising results.
AB - Social media platforms in the last few years have facilitated the development of communities that discuss real-world events, and have shaped the way users interact. The content generated in these platforms reflect a variety of intentions, ranging from social interaction to commercial interest, among many others. The present study aims at the implementation of an automatic intent classification system for a Chilean electricity company social media account. The dataset was created from 5000 tweets that were manually classified by 5 people. If discrepancies were detected, a majority voting scheme was used in order to tag the tweets’ intentions. In order to perform the experimental validation of the automatic classification with the machine learning algorithms, several text representations were used (tf-idf, tf-rfl and bin-rfl). The results obtained from the various tests that were conducted yielded satisfactory results. We also analyzed how to assign automatic responses to frequently asked questions, and obtained promising results.
KW - Intent classification
KW - Machine learning
KW - Social media
UR - http://www.scopus.com/inward/record.url?scp=85025119311&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-58562-8_21
DO - 10.1007/978-3-319-58562-8_21
M3 - Conference contribution
AN - SCOPUS:85025119311
SN - 9783319585611
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 258
EP - 274
BT - Social Computing and Social Media
A2 - Meiselwitz, Gabriele
PB - Springer Verlag
T2 - 9th International Conference on Social Computing and Social Media, SCSM 2017, held as part of the 19th International Conference on Human-Computer Interaction, HCI 2017
Y2 - 9 July 2017 through 14 July 2017
ER -