<![CDATA[Brand exposure monitoring is a specialized form of media monitoring that aims to monitor brand exposure and brand mentions in publicity media. This research proposes a framework model for monitoring brand exposure through analysis of publicity media utilizing query formulation and topic modeling, aimed at extracting important topics contained in external media or mass media, and comparing them with the institution's internal media. Topic extraction is performed using Latent Dirichlet Allocation (LDA) method on text data in the form of SERP (Search Engine Results Page) snippets. Subsequent processing utilizes Jaccard index and cosine similarity calculations. The output of the framework includes visualizations of topics representing themes of articles exposing the institution's brand, and measurement metrics that are useful for decision-making analysis related to media management and institutional communication. Testing the proposed framework using data samples resulted in expected output, namely the topic groups formed can represent dominant topics in both media categories in a given monitoring period, with the highest coverage rate reaching 82.69% and similarity of 33.97%. The use of LDA method in this study does have its limitations, specifically the formation of topic groups that do not purely contain a singular topic, but rather consist of several subtopics. However, this does not diminish the usefulness of the framework.]]>
Copyrights © 2024
