A State College Customer Feedback Data Analysis using Machine Learning-Based Algorithm

Obtaining customers’ view through text-based feedback in a survey is considered an important process for organizations including education sector since it provides an overview of the different relevant aspects which aid administrators in planning, policy making and decision making. Over the years, academic institutions have collected vast amounts of textual data through survey. However, analyzing voluminous amounts of unstructured customer feedback to gain a general view of their concerns and sentiments remains a challenge for the institution. This study conducted a text analysis of the feedbacks from the customer satisfaction survey of one of the State Colleges in Region V, Philippines for academic year 2018-2019. A machine learning-based algorithm such as topic modeling using Latent Dirichlet Allocation (LDA) was employed in the study for the automatic summarization of text and extraction of topics from these unstructured data. Moreover, it describes the text mining process performed to retrieve useful information from the huge amount of text-based data. The topmost concerns extracted from the customer feedback were then identified. In the result, specific concerns for offices were revealed such as staffing, environment, customer feedback system, and IT system. Furthermore, issues on security personnel and student assistants’ attitude as well as library operation and management are notably highlighted in the feedback.

Keywords – Latent Dirichlet Allocation (LDA), machine learning-based algorithm, topic modeling