Automatic filtering non-English complaint emails in tourism industry using N-gram extraction and classification techniques
Todsanai Chumwatana, Ichayaporn Chuaychoo · 2016
Over the decades, tourism is a rapidly growing industry and has become one of the largest economic sectors in the world which leads to the intensive competition among tourism businesses. Most businesses seek for effective strategies to increase a scale of customer base from both local and international markets by focusing on customer satisfaction as it is a key indicator of customer repurchase intention and loyalty. Due to this reason, businesses always provide customers the communication channels such as e-mail in order to perceive customer experiences. Consequently, large number of emails has been sent to the businesses. However, extracting complaint emails from huge amount of emails with different languages is very challenging. This paper proposes a novel methodology to automatically extract complaint emails from inbox for any language. N-gram extraction, which is language independent, and classification technique have been applied in the proposed technique. The experimental results show that the classification performance of the proposed technique is approximately 90%. This shows significant advantages for tourism businesses to immediately respond to customers who have negative experiences, in order to retain goodrelationship and gain larger number of customers.