A carregar...
Publicação
Graph neural networks in e-commerce: unveiling competitive dynamics through search engine´s keywords analysis
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.9 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
This thesis presents a novel framework for e-commerce competitive analysis using Graph
Neural Networks (GNNs), with a focus on Search Engine Optimization (SEO). Through
advanced machine learning techniques, including dimensionality reduction and clustering, it
addresses the challenge of identifying competitors within the dynamic SEO landscape. The
study pioneers the use of a custom GraphSAGE model and proposes a Competitor Retrieval
Algorithm alongside a Keyword Recommendation System. These tools not only refine e commerce strategy but also unveil intricate market structures and relationships. While
acknowledging data limitations and the absence of true labels, the research underscores the
potential of GNNs in e-commerce, offering a methodological base for future exploration and
continuous innovation in machine learning and marketing analytics.
Descrição
Palavras-chave
Graph neural networks (Gnns) Network analysis Competitive landscape analysis Knowledge graph Keyword distance Fasttext Embeddings E-commerce Graphsage Bipartite networks