Publications

Learning Product Representations for Generating Reviews for Cold Products

Fatemeh Pourgholamali and Mohsen Kahani and Zeinab Noorian and Ebrahim Bagheri
Reference:
Fatemeh Pourgholamali; Mohsen Kahani; Zeinab Noorian and Ebrahim Bagheri Learning Product Representations for Generating Reviews for Cold Products. In Knowledge-based Systems, 2021.
Links to Publication: [www][pdf]
Abstract:
Existing work in the literature have shown that the number and quality of product ratings andreviews have a direct correlation with the product purchase rates in online e-commerce portals.However, the majority of the products on e-commerce portals do not have any ratings or reviewsand are known as cold products (∼90% of products on Amazon are cold). As such, there has beengrowing interest in generating reviews for cold products by selectively transferring reviews fromother similar yet warm products. Our work in this paper focuses on this specific problem and gen-erates reviews for cold products through review selection. Similar to existing work in the literature,our work assumes a relationship between product attribute-values and the reviews that productsreceive. However, unlike the literature, our method (1) is not restricted to the exact surface formof a product attribute name; and, (2) can distinguish between the same attribute expressed in dif-ferent forms. We achieve these two important characteristics by proposing methods to learn neuralproduct representations that capture the semantics of product attribute-values as they relate touser reviews. More specifically, our work offers (i) an approach to learn neural representationsof product attribute-values within a shared embedding space as product reviews; (ii) a weightedcomposition strategy to develop product representations from the representation of its attributes;and, (iii) a review selection method that selects relevant reviews for the composed product repre-sentation within the neural embedding space. We show through our extensive experiments on fivedatasets consisting of products fromCNET.comand movies fromrottentomatoes.comthat ourmethod is able to show stronger performance compared to several baselines on ROUGE-2 metrics.
Bibtex Entry:
@article{kbs21, title = {Learning Product Representations for Generating Reviews for Cold Products}, journal = {Knowledge-based Systems}, author = {Fatemeh Pourgholamali and Mohsen Kahani and Zeinab Noorian and Ebrahim Bagheri}, abstract = {Existing work in the literature have shown that the number and quality of product ratings andreviews have a direct correlation with the product purchase rates in online e-commerce portals.However, the majority of the products on e-commerce portals do not have any ratings or reviewsand are known as cold products (∼90% of products on Amazon are cold). As such, there has beengrowing interest in generating reviews for cold products by selectively transferring reviews fromother similar yet warm products. Our work in this paper focuses on this specific problem and gen-erates reviews for cold products through review selection. Similar to existing work in the literature,our work assumes a relationship between product attribute-values and the reviews that productsreceive. However, unlike the literature, our method (1) is not restricted to the exact surface formof a product attribute name; and, (2) can distinguish between the same attribute expressed in dif-ferent forms. We achieve these two important characteristics by proposing methods to learn neuralproduct representations that capture the semantics of product attribute-values as they relate touser reviews. More specifically, our work offers (i) an approach to learn neural representationsof product attribute-values within a shared embedding space as product reviews; (ii) a weightedcomposition strategy to develop product representations from the representation of its attributes;and, (iii) a review selection method that selects relevant reviews for the composed product repre-sentation within the neural embedding space. We show through our extensive experiments on fivedatasets consisting of products fromCNET.comand movies fromrottentomatoes.comthat ourmethod is able to show stronger performance compared to several baselines on ROUGE-2 metrics.}, year = {2021}, webpdf = {http://ls3.rnet.ryerson.ca/wiki/images/7/75/Fatemeh-kbs2021.pdf}, url = {https://www.journals.elsevier.com/knowledge-based-systems} }




Powered by WordPress