Institutional Repository of Vilnius University
Mišrios rinkodaros modelių taikymas e-komercijos įmonėje /
Abstract
dc:descriptionE-commerce is a rapidly expanding field, and assessing marketing effectiveness is crucial for achieving optimal outcomes. Marketing mix models have emerged as a reliable and commonly utilized tool for this purpose. This thesis aims to explore and evaluate the application of marketing mix models within a specific e-commerce company. The primary objectives are to identify the appropriate functional form of the model, effectively transform and select data, construct ordinary least squares and regularized (Lasso, Ridge, and Elastic Net) models, and conduct a comprehensive evaluation and comparison. The study compares four modeling approaches and three sets of variables. Findings indicate that Ridge regression was not suitable for the dataset examined. The Elastic Net and Lasso models demonstrated close resemblance to each other and exhibited similarities to ordinary least squares models based on goodness-of-fit criteria and decomposition plots. The models developed using these three methods exhibited correctness, demonstrated favorable goodness-of-fit metrics, and aligned logically with the decomposition plots. When comparing models with different variable groups, it was evident that those with a greater number of variables outperformed others, exhibiting superior goodness-of-fit values, accurate campaign estimation, and more stable model coefficients. Ultimately, it was concluded that no single model stood out as the best, but potential candidates included Lasso, Elastic Net, or OLS models with greater number of variables.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Užpelkis, Antanas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192845810&prefLang=en_US