Strategies for Improving the Quality of Logistics Courier Services Through Priority Problem-solving Based on Multiclass Classification

PT Tiki Jalur Nugraha Ekakurir (JNE) is the most popular courier service in the country and ranked the first in the category of courier service companies in Indonesia Top Brand 2018. Despite being a top brand, JNE still faces a variety of complaints posted by customers on the internet and social media. This makes it interesting to study as a means of evaluating and finding strategies to improve the quality of logistics services using a sentiment analysis method with classification models of Naive Bayes Classifier and Support Vector Machine as well as multiclass classification. This study aims to review and determine which classification model is the most appropriate for the dataset, find out the rating of sentiments and dimensions in order to measure the quality of logistics company services and identify what problems need to be prioritized. The population in this study are all the tweets mentioned to JNE's Twitter accounts. Tweet samples in this study were collected over a 30-day period, from 1 September 2019 to 30 September 2019, totalling 11,134 tweets. This study revealed that JNE is dominated by negative sentiment of 97.82% and positive sentiment of 2.18%. The main problems (high priority) faced by JNE are the dimensions of Personnel Contact Quality and Timeliness. Then, the dimensions of Information Quality, Order Discrepancy Handling and Order Accuracy are problems with medium priority. Meanwhile, JNE's low priority issues are the dimensions of Order Conditions, Ordering Procedures, Ordering Release Quantities, and Order Quality. Therefore, JNE needs to improve the examination of personal messages and provide a quick response to complaints about customer problems and increase commitment in every delivery service to match the estimated delivery time. It is recommended in future studies to apply the Naive Bayes Classifier classification model in Twitter dataset processing because it has fast, good, and quite precise processing results. Future research can also use the application of ABC analysis which can be used as a basis for determining the number of priority dimensions of a service's problem solving.