Estimasi Strategi Pemasaran Berdasarkan Data Engagement Media Sosial Menggunakan Analisis Intervensi-Kalman Filter

Putri, Novi Saumi (2025) Estimasi Strategi Pemasaran Berdasarkan Data Engagement Media Sosial Menggunakan Analisis Intervensi-Kalman Filter. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002211132-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Perkembangan pesat e-commerce dan teknologi internet memberikan dampak signifikan terhadap pemasaran digital. Berdasarkan hal ini, media sosial menjadi salah satu alat yang efektif untuk meningkatkan engagement pelanggan dan mempengaruhi citra brand. Penelitian ini bertujuan untuk mengembangkan model estimasi strategi pemasaran berbasis data engagement menggunakan Analisis Intervensi-Kalman Filter. Pemodelan ini berfokus pada pengukuran keterlibatan audiens dengan brand e-commerce melalui media sosial Instagram, khususnya pada kampanye Shopee. Dalam proses penelitian, dilakukan analisis time series dengan menggunakan model ARIMA untuk mengestimasi engagement sebelum intervensi, serta penerapan metode Kalman Filter untuk memperbaiki estimasi berdasarkan data terbaru.Hasil penelitian menunjukkan bahwa model ARIMA(0,1,1) merupakan model terbaik dengan nilai AIC dan SBC yang lebih rendah, serta residual yang mendekati white noise dan terdistribusi normal, dibandingkan dengan model lainnya. Penggunaan Analisis Intervensi Kalman Filter memberikan estimasi yang paling akurat, dengan nilai MAPE sebesar 0,20% artinya lebih rendah dibandingkan dengan model ARIMA (9,95%) dan Analisis Intervensi (9,89%). Model ini memberikan gambaran bahwa penggunaan Kalman Filter dapat memperbaiki estimasi dan prediksi yang lebih responsif terhadap perubahan perilaku konsumen.
===============================================================================================================================
The rapid growth of e-commerce and internet technologies has had a profound impact on digital marketing strategies. Among various digital platforms, social media has emerged as an effective tool for enhancing customer engagement and shaping brand image. This study aims to develop a predictive marketing strategy model based on engagement data using Intervention Analysis and the Kalman Filter. The modeling focuses on estimation audience engagement with e-commerce brands via Instagram, specifically analyzing Shopee’s marketing campaigns. Time series analysis was employed, with the ARIMA model used to estimate engagement trends prior to intervention events. The Kalman Filter was subsequently applied to refine these estimates using real-time data updates. The results indicate that the ARIMA(0,1,1) model offers the best performance, exhibiting lower AIC and SBC values and residuals that approximate white noise and follow a normal distribution, outperforming alternative models. The combinedIntervention-Kalman Filter approach yielded the most accurate engagement estimation, with a Mean Absolute Percentage Error (MAPE) of 0.20% which significantly lower than that of the ARIMA model (9.95%) and the Intervention Analysis model alone (9.89%). These findings suggest that incorporating the Kalman Filter into the analysis enhances predictive accuracy and provides a more adaptive model for capturing shifts in consumer behavior.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Strategi Pemasaran, Engagement Media Sosial, Analisis Intervensi, Kalman Filter, Marketing Strategy, Social Media Engagement, Intervention Analysis, Kalman Filter.
Subjects:	Q Science > QA Mathematics > QA402.3 Kalman filtering.
Divisions:	Faculty of Mathematics, Computation, and Data Science > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Novi Saumi Putri
Date Deposited:	01 Aug 2025 02:22
Last Modified:	01 Aug 2025 02:22
URI:	http://repository.its.ac.id/id/eprint/124778

Actions (login required)

View Item