Gyanit Singh

Query suggestion for E-commerce sites

Query suggestion module is an integral part of every search engine. It helps search engine users narrow or broaden their searches. Published work on query suggestion methods has mainly focused on the web domain. But, the module is also popular in the domain of e-commerce for product search. In this paper, we discuss query suggestion and its methodologies in the context of e-commerce search engines. We show that dynamic inventory combined with long and sparse tail of query distribution poses unique challenges to build a query suggestion method for an e-commerce marketplace. We compare and contrast the design of a query suggestion system for web search engines and e-commerce search engines. Further, we discuss interesting measures to quantify the effectiveness of our query suggestion methodologies. We also describe the learning gained from exposing our query suggestion module to a vibrant community of millions of users.

Rewriting null e-commerce queries to recommend products

In e-commerce applications product descriptions are often concise. E-Commerce search engines often have to deal with queries that cannot be easily matched to product inventory resulting in zero recall or null query situations. Null queries arise from differences in buyer and seller vocabulary or from the transient nature of products. In this paper, we describe a system that rewrites null e-commerce queries to find matching products as close to the original query as possible. The system uses query relaxation to rewrite null queries in order to match products. Using eBay as an example of a dynamic marketplace, we show how using temporal feedback that respects product category structure using the repository of expired products, we improve the quality of recommended results. The system is scalable and can be run in a high volume setting. We show through our experiments that high quality product recommendations for more than 25% of null queries are achievable.

User behavior in zero-recall ecommerce queries

User expectation and experience for web search and eCommerce (product) search are quite different. Product descriptions are concise as compared to typical web documents. User expectation is more specific to find the right product. The difference in the publisher and searcher vocabulary (in case of product search the seller and the buyer vocabulary) combined with the fact that there are fewer products to search over than web documents result in observable numbers of searches that return no results (zero recall searches). In this paper we describe a study of zero recall searches. Our study is focused on eCommerce search and uses data from a leading eCommerce sites user click stream logs. There are 3 main contributions of our study: 1) The cause of zero recall searches; 2) A study of users reaction and recovery from zero recall; 3) A study of differences in behavior of power users versus novice users to zero recall searches.

A Study of Query Term Deletion Using Large-Scale E-commerce Search Logs

Query term deletion is one of the commonly used strategies for query rewriting. In this paper, we study the problem of query term deletion using large-scale e-commerce search logs. Specifically, we focus on queries that do not lead to user clicks and aim to predict a reduced and better query that can lead to clicks by term deletion. Accurate prediction of term deletion can potentially help users recover from poor search results and improve shopping experience. To achieve this, we use various term-dependent and query-dependent measures as features and build a classifier to predict which term is the most likely to be deleted from a given query. Our approach is data-driven. We investigate the large-scale query history and the document collection, verify the usefulness of previously proposed features, and also propose to incorporate the query category information into the term deletion predictors. We observe that training within-category classifiers can result in much better performance than training a unified classifier. We validate our approach using a large collection of query sessions logs from a leading e-commerce site and demonstrate that our approach provides promising performance in query term deletion prediction.

Ranking approach to RecSys Challenge

In this paper we describe our approach to solve RecSys Challenge 2014. The challenge is to rank users auto generated IMDB rating tweets by their favorited and shared count. Our approach is to formulate this as a ranking problem. We treat a single user as a query and all of the known tweets are treated as matching documents. We then apply various learning to rank approaches and pick the best performing.

Query Suggestion with Large Scale Data

Abstract Explosive growth of information has created a challenge for search engines in various domains to handle large scale data. It is still difficult for search engines to fully understand user intent in many scenarios. To address this, most search engines provide assistive features to the user which help the users in managing their information need. Query Suggestion (Related Searches) is one such feature which is an integral part of all search engines. It helps steer users toward queries which are more likely to help them succeed in their search missions. There has been extensive research in this field. In this chapter, we discuss state-of-the-art techniques to build a Query Suggestion system. Specifically, we describe the strengths and limitations of different approaches. We also describe salient characteristics of large scale data sets like query corpora and click-stream logs. We walk the reader through the design, implementation, and evaluation of large scale Query Suggestion systems in practice. We show how challenges related to sparsity in the long tail, biases in user data, and speed of algorithms can be tackled at industry scale.

Blurb Mining: Discovering Interesting Excerpts from E-commerce Product Reviews

Product reviews on modern e-commerce websites have evolved into repositories of valuable firsthand opinions on products. Showcasing the opinions that reviewers express on a product in a succinct way can not only promote the product, but also provide an engaging experience that simplifies the shopping journey for online shoppers. In the case of traditional media such as movies and books, employingblurbs or excerpts from critic reviews for promotional purposes is an established practice among movie publicists and book editors that has proven to be an effective way of capturing attention of customers. Such excerpts can be discovered from e-commerce product reviews to highlight interesting reviewer opinions and add emotive elements to otherwise bland e-commerce product pages. While traditional movie or book blurbs are manually extracted, they must be automatically extracted from e-commerce product reviews owing to the scale of catalogues. Further, traditional blurbs are generally phrased to be very positive in tone and sometimes may take some words out of context. However, excerpts for e-commerce products must represent the true opinions of the reviewers and must capture the context in which the words were used to retain trust of users. To that end, we introduce the problem of extracting engaging excerpts from e-commerce product reviews in this paper. We present methods to automatically discover such excerpts from reviews at scale by leveraging natural language properties such as syntactic structure of sentences and sentiment, and discuss some of the underlying challenges. We further present an evaluation of the effectiveness of the proposed methods in terms of the quality of the blurbs generated and their ranking orders produced.