Finding the Right Influencers on Instagram for Endorsement Product Using Text Mining

Hedy Pamungkas, Riswan Haryo Yudhianto, Bern Jonathan Sembiring, Indra Budi

Research output: Contribution to journalArticlepeer-review


The growth of the Internet has significantly changed our way of life. Social media is being used not only for getting connected with people but also for media sharing, diffusing information, and even marketing. Social media marketing has a significant impact because it can target campaigns appropriately based on the segmentation of the products to be marketed. Furthermore, this approach can provide campaign costs that are less expensive than other traditional methods. One of the methods which are generally used is an endorsement. With the popularity of social media, a new type of endorsement called social media endorsement has been born. This type of endorsement incorporates social media influencers to promote products and goods. However, product and brand owners find it difficult to select suitable influencers to endorse their products. Therefore, we try to solve this problem by creating a recommender system using the clustering method. We collect data from Instagram, one of the most popular social media platforms. The result showed that three clusters produced out of the data had high quality. In addition, we applied Silhouette Coefficient to validate the result, which produced a positive result on 7.277 X 10-3. With such a result, we conclude that this model could be used to categorize which brands or products an influencer normally endorse and recommend product or brand owners if an influencer is suitable based on clustering.

Original languageEnglish
Pages (from-to)1395-1402
Number of pages8
JournalInternational Journal on Advanced Science, Engineering and Information Technology
Issue number4
Publication statusPublished - 2022


  • Clustering
  • Endorsement
  • Influencer
  • Instagram
  • Marketing.


Dive into the research topics of 'Finding the Right Influencers on Instagram for Endorsement Product Using Text Mining'. Together they form a unique fingerprint.

Cite this