Prediction is very hard, especially about conversion. Predicting user\n purchases from clickstream data in fashion e-commerce

2019/06/30 by Luca Bigon, Bigon, Luca, Giovanni Cassani +11 · 1 citation
Business, Management and Accounting · Social Sciences · Computer Science · #Consumer Market Behavior and Pricing #Digital Marketing and Social Media #Recommender Systems and Techniques

paper · pdf · doi:10.48550/arxiv.1907.00400

Abstract

Knowing if a user is a buyer vs window shopper solely based on clickstream\ndata is of crucial importance for ecommerce platforms seeking to implement\nreal-time accurate NBA (next best action) policies. However, due to the low\nfrequency of conversion events and the noisiness of browsing data, classifying\nuser sessions is very challenging. In this paper, we address the clickstream\nclassification problem in the fashion industry and present three major\ncontributions to the burgeoning field of AI in fashion: first, we collected,\nnormalized and prepared a novel dataset of live shopping sessions from a major\nEuropean e-commerce fashion website; second, we use the dataset to test in a\ncontrolled environment strong baselines and SOTA models from the literature;\nfinally, we propose a new discriminative neural model that outperforms neural\narchitectures recently proposed at Rakuten labs.\n

Cited by

• SIGIR 2021 E-Commerce Workshop Data Challenge

Related

• Predicting e-commerce customer conversion from minimal temporal patterns on symbolized clickstream trajectories
• Predicting next shopping stage using Google Analytics data for E-commerce applications
• Predicting purchasing intent: Automatic Feature Learning using Recurrent Neural Networks
• Predicting online user behaviour using deep learning algorithms
• Neural Search: Learning Query and Product Representations in Fashion E-commerce