A carregar...
Publicação
An OLX Motors case: comprehend user behaviour to shape recommendations
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.25 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
The analysis of online user behaviour can shape the way a business like OLX Motors recommends cars to potential buyers. This paper focuses on different ways of getting insights from the searches and the interactions of the users on the websites. Understanding first how users select filters and how long their searches are, led to a first AB test to check whether suggesting filters to select could or not help the user experience. The results showed that looking at different models in a search could bring users closer to their car of interest. Intuitively, recommending car models would eventually enhance the user's journey. This evolved into utilizing word2vec, a widely used algorithm to explore relationships between words in sentences, to draw similarities between similar cars. The data used is the sequence of users' search sessions and not the technical characteristics of the cars. With this a clustering technique was performed to individualize groups of similar vehicles.
Descrição
Internship Report presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Business Analytics
Palavras-chave
Recommender Systems Machine Learning Cars Market User Behaviour AB Test