Popularity Estimation and New Bundle Generation using Content and Context based Embeddings

TL;DR

The paper tackles static bundle generation by introducing a set of novel bundle popularity metrics and leveraging content- and context-aware embeddings to sample and optimize bundles, demonstrated on Steam data. It defines six metrics (P_eb, P_mb, N^0_b, P^B_b, D_b, C^B) and uses SBERT/FastText sentence embeddings along with Prod2Vec/MetaProd2Vec context embeddings to predict bundle popularity with a logistic regression model. Through four greedy sampling strategies, the approach upgrades existing bundles and generates new bundles that outperform baselines by substantial margins across multiple popularity metrics, validating the method's effectiveness and generality. The work contributes a holistic framework for bundle quality assessment and generation, with practical implications for static bundle design in large catalogs, while outlining avenues for dynamic bundling and broader domain validation.

Abstract

Recommender systems create enormous value for businesses and their consumers. They increase revenue for businesses while improving the consumer experience by recommending relevant products amidst huge product base. Product bundling is an exciting development in the field of product recommendations. It aims at generating new bundles and recommending exciting and relevant bundles to their consumers. Unlike traditional recommender systems that recommend single items to consumers, product bundling aims at targeting a bundle, or a set of items, to the consumers. While bundle recommendation has attracted significant research interest recently, extant literature on bundle generation is scarce. Moreover, metrics to identify if a bundle is popular or not is not well studied. In this work, we aim to fulfill this gap by introducing new bundle popularity metrics based on sales, consumer experience and item diversity in a bundle. We use these metrics in the methodology proposed in this paper to generate new bundles for mobile games using content aware and context aware embeddings. We use opensource Steam Games dataset for our analysis. Our experiments indicate that we can generate new bundles that can outperform the existing bundles on the popularity metrics by 32% - 44%. Our experiments are computationally efficient and the proposed methodology is generic that can be extended to other bundling problems e.g. product bundling, music bundling.

Figures (10)

• Figure 1: Example of a Bundle in Steam Games
• Figure 2: Game Features in $Large$ Dataset
• Figure 3: Intersection of Games in Bundles
• Figure 4: Distribution of Purchases in $Small$ Dataset
• Figure 5: Distribution of Purchases in $Large$ dataset