Recommendations Using Information from Multiple Association Rules: A Probabilistic Approach

Published on Sep 1, 2015in Information Systems Research
· DOI :10.1287/ISRE.2015.0583
Abhijeet Ghoshal3
Estimated H-index: 3
(University of Louisville),
Syam Menon11
Estimated H-index: 11
(UTD: University of Texas at Dallas),
Sumit Sarkar21
Estimated H-index: 21
(UTD: University of Texas at Dallas)
Business analytics has evolved from being a novelty used by a select few to an accepted facet of conducting business. Recommender systems form a critical component of the business analytics toolkit and, by enabling firms to effectively target customers with products and services, are helping alter the e-commerce landscape. A variety of methods exist for providing recommendations, with collaborative filtering, matrix factorization, and association-rule-based methods being the most common. In this paper, we propose a method to improve the quality of recommendations made using association rules. This is accomplished by combining rules when possible and stands apart from existing rule-combination methods in that it is strongly grounded in probability theory. Combining rules requires the identification of the best combination of rules from the many combinations that might exist, and we use a maximum-likelihood framework to compare alternative combinations. Because it is impractical to apply the maximum likelihood framework directly in real time, we show that this problem can equivalently be represented as a set partitioning problem by translating it into an information theoretic context-the best solution corresponds to the set of rules that leads to the highest sum of mutual information associated with the rules. Through a variety of experiments that evaluate the quality of recommendations made using the proposed approach, we show that i a greedy heuristic used to solve the maximum likelihood estimation problem is very effective, providing results comparable to those from using the optimal set partitioning solution; ii the recommendations made by our approach are more accurate than those made by a variety of state-of-the-art benchmarks, including collaborative filtering and matrix factorization; and iii the recommendations can be made in a fraction of a second on a desktop computer, making it practical to use in real-world applications.
  • References (50)
  • Citations (8)
📖 Papers frequently viewed together
6 Citations
1 Author (Jiye Li)
14 Citations
78% of Scinapse members use related papers. After signing in, all features are FREE.
#1Kartik Hosanagar (UPenn: University of Pennsylvania)H-Index: 20
#2Daniel M. Fleder (UPenn: University of Pennsylvania)H-Index: 6
Last. Andreas Buja (UPenn: University of Pennsylvania)H-Index: 36
view all 4 authors...
Personalization is becoming ubiquitous on the World Wide Web. Such systems use statistical techniques to infer a customer’s preferences and recommend content best suited to him (e.g., “Customers who liked this also liked…”). A debate has emerged as to whether personalization has drawbacks. By making the web hyper-specific to our interests, does it fragment internet users, reducing shared experiences and narrowing media consumption? We study whether personalization is in fact fragmenting the onli...
11 CitationsSource
#1Kartik HosanagarH-Index: 20
#2Daniel M. FlederH-Index: 6
Last. Andreas BujaH-Index: 36
view all 4 authors...
Personalization is becoming ubiquitous on the World Wide Web. Such systems use statistical techniques to infer a customer's preferences and recommend content best suited to him (e.g., “Customers who liked this also liked...”). A debate has emerged as to whether personalization has drawbacks. By making the Web hyperspecific to our interests, does it fragment Internet users, reducing shared experiences and narrowing media consumption? We study whether personalization is in fact fragmenting the onl...
72 CitationsSource
#1Toon Calders (TU/e: Eindhoven University of Technology)H-Index: 31
#2Nele Dexters (University of Antwerp)H-Index: 5
Last. Bart Goethals (University of Antwerp)H-Index: 33
view all 4 authors...
Mining frequent itemsets in a datastream proves to be a difficult problem, as itemsets arrive in rapid succession and storing parts of the stream is typically impossible. Nonetheless, it has many useful applications; e.g., opinion and sentiment analysis from social networks. Current stream mining algorithms are based on approximations. In earlier work, mining frequent items in a stream under the max-frequency measure proved to be effective for items. In this paper, we extended our work from item...
40 CitationsSource
#1Cheng Zhou (National University of Defense Technology)H-Index: 4
#2Boris Cule (University of Antwerp)H-Index: 10
Last. Bart Goethals (University of Antwerp)H-Index: 33
view all 3 authors...
Sequence classification is an important task in data mining. We address the problem of sequence classification using rules composed of interesting itemsets found in a dataset of labelled sequences and accompanying class labels. We measure the interestingness of an itemset in a given class of sequences by combining the cohesion and the support of the itemset. We use the discovered itemsets to generate confident classification rules, and present two different ways of building a classifier. The fir...
9 CitationsSource
Oct 23, 2011 in RecSys (Conference on Recommender Systems)
#1Michael D. Ekstrand (UMN: University of Minnesota)H-Index: 16
#2Michael Ludwig (UMN: University of Minnesota)H-Index: 4
Last. John Riedl (UMN: University of Minnesota)H-Index: 60
view all 4 authors...
Recommender systems research is being slowed by the difficulty of replicating and comparing research results. Published research uses various experimental methodologies and metrics that are difficult to compare. It also often fails to sufficiently document the details of proposed algorithms or the evaluations employed. Researchers waste time reimplementing well-known algorithms, and the new implementations may miss key details from the original algorithm or its subsequent refinements. When propo...
111 CitationsSource
#1Gediminas Adomavicius (UMN: University of Minnesota)H-Index: 28
#2Alexander Tljzhilin (NYU: New York University)H-Index: 40
Last. Rong Zheng (HKUST: Hong Kong University of Science and Technology)H-Index: 35
view all 3 authors...
Initially popularized by, recommendation technologies have become widespread over the past several years. However, the types of recommendations available to the users in these recommender systems are typically determined by the vendor and therefore are not flexible. In this paper, we address this problem by presenting the recommendation query language REQUEST that allows users to customize recommendations by formulating them in the ways satisfying personalized needs of the users. REQU...
42 CitationsSource
#1Bhavik PathakH-Index: 7
#2Robert Garfinkel (UConn: University of Connecticut)H-Index: 19
Last. Fang Yin (UO: University of Oregon)H-Index: 10
view all 5 authors...
Online retailers are increasingly using information technologies to provide value-added services to customers. Prominent examples of these services are online recommender systems and consumer feedback mechanisms, both of which serve to reduce consumer search costs and uncertainty associated with the purchase of unfamiliar products. The central question we address is how recommender systems affect sales. We take into consideration the interaction among recommendations, sales, and price. We then d...
121 CitationsSource
#1Dietmar JannachH-Index: 36
#2Markus ZankerH-Index: 27
Last. Gerhard FriedrichH-Index: 31
view all 4 authors...
In this age of information overload, people use a variety of strategies to make choices about what to buy, how to spend their leisure time, and even whom to date. Recommender systems automate some of these strategies with the goal of providing affordable, personal, and high-quality recommendations. This book offers an overview of approaches to developing state-of-the-art recommender systems. The authors present current algorithmic approaches for generating personalized buying proposals, such as ...
712 Citations
Self-sufficient itemsets are those whose frequency cannot be explained solely by the frequency of either their subsets or of their supersets. We argue that itemsets that are not self-sufficient will often be of little interest to the data analyst, as their frequency should be expected once that of the itemsets on which their frequency depends is known. We present tests for statistically sound discovery of self-sufficient itemsets, and computational techniques that allow those tests to be applied...
65 CitationsSource
#1Yehuda Koren (Yahoo!)H-Index: 45
#2Robert M. BellH-Index: 49
Last. Chris VolinskyH-Index: 26
view all 3 authors...
As the Netflix Prize competition has demonstrated, matrix factorization models are superior to classic nearest neighbor techniques for producing product recommendations, allowing the incorporation of additional information such as implicit feedback, temporal effects, and confidence levels.
4,513 CitationsSource
Cited By8
#1Zhiang Wu (Nanjing Audit University)
#2Changsheng Li (Nanjing University of Finance and Economics)H-Index: 1
Last. Yong Ge (UA: University of Arizona)H-Index: 25
view all 4 authors...
#1Qian Wang (SYSU: Sun Yat-sen University)H-Index: 1
#2Jijun Yu (SYSU: Sun Yat-sen University)H-Index: 1
Last. Weiwei Deng (CityU: City University of Hong Kong)H-Index: 1
view all 3 authors...
The effectiveness of product recommendations is previously assessed based on recommendation accuracy. Recently, individual diversity and aggregate diversity of product recommendations have been recognized as important dimensions in evaluating the recommendation effectiveness. However, the gain of either diversity is usually at the cost of accuracy and the increase of one diversity does not guarantee a significant improvement in the other. A few attempts have been made to achieve reasonable trade...
3 CitationsSource
#1Edianny Carballo-Cruz (University of Ciego de Ávila)
#2Raciel Yera (University of Ciego de Ávila)
Last. Maria Elena Betancourt (UC: University of Camagüey)
view all 4 authors...
Touristic innovation is understood as the implementation of new or improved products, processes, organization and marketing practices, that consider the non-lineal behavior and the knowledge generation. Although the related literature globally establishes the activities that should be contained in an innovation procedure to obtain new and improved products, there are not definitive guidelines regarding the order that should follow the execution of this activities. The current research fills this...
#1Dan Zhang (Sichuan University)
#2Xinrui Wang (Sichuan University)
Last. Changzheng He (Sichuan University)
view all 5 authors...
Rapid growth in the number of smart phones users brings endless opportunities and enormous challenges for mobile Internet applications. How to recommend applications in which users are interested draws attentions of the application stores. In this paper, the features of different categories applications are extracted by analyzing users behavior of downloading. Based on association rules, the recommendations of overall and different categories applications can be generated respectively, as well a...
#1Changsheng Li (Nanjing University of Finance and Economics)H-Index: 1
#2Weichao Liang (Nanjing University of Science and Technology)H-Index: 1
Last. Jie Cao (Nanjing University of Science and Technology)H-Index: 25
view all 4 authors...
The association-rule-based recommendation model is one of the most widely used commercial recommendation engines in e-commerce websites. Existing studies mostly focus on how to select eligible rules to enhance the recommendation performance, but the efficiency of recommendation has been paid few attentions. To remedy this, this paper develops a distributed-computing framework for improving the computational efficiency of rule-based recommendation. Specifically, a tree-typed structure called Orde...
1 CitationsSource
#1Mingxin Gan (USTB: University of Science and Technology Beijing)H-Index: 6
#2Rui Jiang (THU: Tsinghua University)H-Index: 41
Abstract In big data era people are dependent on a variety of social media to manage their social circles. Many online social networks employ social recommendation as an increasingly important component. Although global and local recommendation methods have achieved remarkable success, current studies seldom consider to play advantages of both in social networks. To demonstrate the effectiveness of incorporating local methods to global ones, we first investigated associations between triangular ...
3 CitationsSource
#1Stella Pachidi (University of Cambridge)H-Index: 4
#2Marleen HuysmanH-Index: 25
2 CitationsSource
#1Arun Rai (J. Mack Robinson College of Business)H-Index: 49
1 Citations