Chaniago, M Fajri Davyza (2023) PREDIKSI POPULARITAS FOTO PADA MEDIA SOSIAL INSTAGRAM VIA ANALISIS REGRESI. Other thesis, Institut Teknologi Sepuluh November.
Text
05111940000180_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 3 October 2025. Download (36MB) | Request a copy |
Abstract
Instagram merupakan salah satu dampak dari perkembangan teknologi yang terjadi saat ini. Perhatian yang sering kali diberikan oleh masyarakat pada media sosial Instagram yaitu jumlah ‘like’. Hal ini digunakan untuk menambah exposure dari orang tersebut. Namun, terkadang masyarakat sulit mengetahui berapa jumlah ‘like’ yang didapatkan. Prediksi jumlah ‘like’ pada konten unggahan foto sebenarnya dapat diprediksi dengan menggunakan metode dalam machine learning yaitu dengan menggunakan analisis regresi. Tahapan metode machine learning yang digunakan pada penelitian ini yaitu tahapan pre-process data, kemudian penentuan fitur untuk pemodelan yang digunakan sebagai prediksi, setelah itu dilakukan tahap pembelajaran terhadap pola data, selanjutnya dari pembelajaran pola data dilakukan pembuatan model dan terakhir dilakukan prediksi dari model yang sudah didapatkan. Penelitan ini menggunakan analisis regresi untuk memprediksi jumlah ‘like’ dengan menggunakan sepuluh Influencers Category yaitu ‘Creators & Celebrities’, ‘Digital Creator’, ‘Entrepreneur’, ‘Fashion Model’, ‘General Interest’, ‘Health & Beauty’, ‘Home Services’, ‘Lifestyle Services’, ‘Personal Goods & General Merchandise Stores’, dan ‘Publishers’. Metode yang digunakan dalam penelitian ini yaitu Decision Tree, Random Forest Tree, Catboost, Support Vector Regressor, dan Artificial Neural Network. Dari lima metode yang digunakan, metode Catboost merupakan metode terbaik dengan nilai Mean Square Error (MSE) sebesar 0,55, Root Mean Square Error (RMSE) sebesar 0,74, dan Mean Absolute Error (MAE) sebesar 0,53. Sementara itu, nilai Spearman Correlation yang didapatkan yaitu sebesar 0,95.
=============================================================================================================================
Instagram is one of the impacts of current technological advancements. The attention often given by society on the social media platform Instagram is the number of 'likes'. This is used to increase exposure for the individual. However, sometimes people find it difficult to determine the number of 'likes' received. The prediction of the number of 'likes' on photo content can actually be achieved using a machine learning method called regression analysis. The stages of the machine learning method used in this research are as follows: data pre-processing, determining features for modeling used in prediction, followed by the learning stage to identify data patterns. Subsequently, a model is created based on the learned data patterns, and finally, predictions are made using the obtained model. This research employs regression analysis to predict the number of 'likes', using ten Influencers Categories: 'Creators & Celebrities', 'Digital Creator', 'Entrepreneur', 'Fashion Model', 'General Interest', 'Health & Beauty', 'Home Services', 'Lifestyle Services', 'Personal Goods & General Merchandise Stores', and 'Publishers'. The methods used in this study are Decision Tree, Random Forest Tree, Catboost, Support Vector Regressor, and Artificial Neural Network. Out of the five methods used, Catboost proves to be the best-performing method, with a Mean Square Error (MSE) value of 0.55, Root Mean Square Error (RMSE) value of 0.74, and Mean Absolute Error (MAE) value of 0.53. Additionally, the obtained Spearman Correlation value is 0.95.
Keywords: Instagram, Like, Popularity, Regression.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Instagram, Like, Popularitas, Prediksi, Regresi. Instagram, Like, Popularity, Regression. |
Subjects: | T Technology > T Technology (General) > T57.5 Data Processing |
Divisions: | Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis |
Depositing User: | M.Fajri Davyza Chaniago |
Date Deposited: | 03 Aug 2023 07:46 |
Last Modified: | 03 Aug 2023 07:46 |
URI: | http://repository.its.ac.id/id/eprint/101548 |
Actions (login required)
View Item |