Institutional Repository of Vilnius University
„Amazon” elektroninės parduotuvės prekių pardavimų reitingo prognozavimas paremtas jų skelbimų duomenimis /
Abstract
dc:descriptionIn online stores, millions of sellers and even more products compete, so in such a competition it is necessary to develop the best product. This study will use the data of one seller's products uploaded to the Amazon platform from the "shower vanities and accessories" category, describing the features of the products and their listings. It is proposed to consider the popularity of products as the sales rating given by Amazon to the product and to evaluate it according to the parameters of the product and its listing. Using the k-means method, three groups of similar products are formed - clusterized. In each cluster, linear, Lasso and random forest regression models are trained, where the dependent variable is the sales rating, independent - product's, its manufacturer's and advertisement attributes. After training and testing the algorithms, it was found that linear regression models are not suitable for evaluation, random forests were superior in every case. It was found that clustering did not bring any additional accuracy.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Amšiejus, Matas,
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:ELABAETD192844503&prefLang=en_US