Real-time Recommendation System Research Articles

Online discrete choice models: Applications in personalized recommendations

This paper presents a framework for estimating and updating user preferences in the context of app-based recommender systems. We specifically consider recommender systems which provide personalized menus of options to users. A Hierarchical Bayes procedure is applied in order to account for inter- and intra-consumer heterogeneity, representing random taste variations among individuals and among choice situations (menus) for a given individual, respectively. Three levels of preference parameters are estimated: population-level, individual-level and menu-specific. In the context of a recommender system, the estimation of these parameters is repeated periodically in an offline process in order to account for trends, such as changing market conditions. Furthermore, the individual-level parameters are updated in real-time as users make choices in order to incorporate the latest information from the users. This online update is computationally efficient which makes it feasible to embed it in a real-time recommender system. The estimated individual-level preferences are stored for each user and retrieved as inputs to a menu optimization model in order to provide recommendations. The proposed methodology is applied to both Monte-Carlo and real data. It is observed that the online update of the parameters is successful in improving the parameter estimates in real-time. This framework is relevant to various recommender systems that generate personalized recommendations ranging from transportation to e-commerce and online marketing, but is particularly useful when the attributes of the alternatives vary over time.

Real-Time Advisory System Mitigates Drilling Vibrations in a Lateral Section

This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 188942, “Mitigating Drilling Vibrations in a Lateral Section Using a Real-Time Advisory System,” by J.R. Bailey, SPE, and G.S. Payette, SPE, ExxonMobil; M.T. Prim, J. Molster, and A.W. Al Mheiri, Zakum Development; P.G. McCormack, SPE, Halliburton; and K. LeRoy, Pason Systems, prepared for the 2017 Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, 13–16 November. The paper has not been peer reviewed. A real-time drilling-data analysis and recommendation system that leverages surface drilling data was deployed in a lateral section of a well from an artificial island in Abu Dhabi. A key objective of this technology was to provide the driller with an easy-to-use display of a novel drilling-performance map. This paper presents data from the demonstration run, illustrates the features of this technology, and provides general observations on optimization of drilling parameters in this hole section. Background A research prototype drilling advisory system (DAS) was set up on an island rig with a wellsite information transfer specification (WITS) data feed from the mud-logger cabin over a fiber-to-serial cable to the computer in the driller’s cabin. The DAS helps the driller make decisions for managing controllable drilling parameters to enable a high rate of penetration (ROP) and low vibrations. The system uses surface data only (i.e., data obtained from sensors instrumented on surface equipment) from the mud logger. The platform then analyzes the data and displays basic surveillance information, including heat maps that vary with the controllable drilling parameters. On the basis of this data and processing algorithms, the system makes recommendations for the controllable parameter values. Data from three intervals were analyzed, and two of the three saw contrasting trends of stick/slip [torsional severity estimate (TSE)] and mechanical specific energy (MSE) with increasing rotary speed, whereas one interval did not see such a tradeoff. The complex interplay between these dysfunctions requires an application-specific real-time advisory system to provide optimal drilling performance. The control variables for this drilling application were ROP and rotary speed. Intervals for Data Analysis The 8½-in. drilling interval started from the 9 5/8-in. casing shoe at 13,916 ft, and section total depth was called at 23,974 ft. Drilling data from this interval was subdivided into seven intervals corresponding to drilling tests, with parameter recommendations from the DAS analysis. Examples of Drilling-Parameter Optimization To initiate the DAS, the driller conducts a drilling test by starting the calibration mode on the advisory system to reset the system memory. Parameter set-point values then must be maintained relatively constant for a time sufficient to determine good average values for the drilling response. The advisory system requires a statistically significant sampling of data at stationary parameter values. When several such calibration points have been recorded, the system then provides automated analysis of the drilling test for the span of operating parameters. If the formation changes over this interval substantially, the calibration process should be repeated.

"Real-time recommendations in a multi-domain environment" by Emanuel Lacic with Prateek Jain as coordinator

Emanuel Lacić is a Software Engineer and Scientific Researcher in the Social Computing division at the Know-Center. Additionally, he is a PhD student at the Interactive Systems & Data Science Institute of Graz University of Technology. He has a M.Sc. and B.Sc. in Software Engineering and Information Systems from the Faculty of Electrical Engineering and Computing at the University of Zagreb. His PhD focus is on real-time recommender systems in large-scale settings. Apart from that, his research is utilized in several industry and European-funded projects such as Learning Layers, AFEL or MoreGrasp. His main research interests are in the fields of Recommender Systems, Deep Learning, Information Retrieval, Big Data, as well as Social and Complex Network Analysis. For more information please see https://lacic.github.io/

RecTime: Real-Time recommender system for online broadcasting

Recommender systems for online broadcasting become important as the number of channels has been increasing. In online broadcasting, to provide accurate recommendation, recommender systems should take time factors as well as users’ condition into account, but the conventional systems don’t. This paper proposes a real-time recommender system for online broadcasting called RecTime which considers time factors and preferences simultaneously. Specifically, RecTime employs a 4-d tensor factorization, which considers two more dimensions regarding the time factors, while typical collaboriative filtering methods only consider two dimensions, users and items. By factorizing the 4-d tensor, the system naturally identifies the recommendation time and the items at the same time. In our experiments on real-world data, RecTime properly models users’ watching patterns and significantly outperforms previous methods in terms of the accuracy on the recommendation time as well as the items.

Web Mining: Opinion and Feedback Analysis for Educational Institutions

Big amount of data available in the forms of reviews, opinions, feedbacks, remarks, comments, observations, clarifications, and explanations that require a robust mechanism to store, retrieve, analyze, and management. In this paper, we are proposing system that provides review or summary of above mention text data available on web for an educational institute. Due to big data size above system can be extended as real time recommendation system for Big Data analysis as future enhancement.

The Application of Cloud Based on Latent Factor Algorithm in Personalized Recommendation

Recommendation system is a commercial marketing method. What more, the system could increase adhesion and satisfaction of consumers to the website which brings great commercial benefit to electronic commerce. But with big data ages coming, it makes a great challenge to real-time recommendation system. As for latent factor class collaborative filtering algorithm, a distributed constructed latent factor algorithm based on cloud is presented in this paper. The algorithm could keep collaborative filtering in good recommendation and ensure the real time in massive data environment. The simulation shows that the algorithm could achieve the recommendation efficiently and quickly. High speedup and scalability are proved.

A new heuristic for learning Bayesian networks from limited datasets: a real-time recommendation system application with RFID systems in grocery stores

Bayesian networks (BNs) are a useful tool for applications where dynamic decision-making is involved. However, it is not easy to learn the structure and conditional probability tables of BNs from small datasets. There are many algorithms and heuristics for learning BNs from sparse datasets, but most of these are not concerned with the quality of the learned network in the context of a specific application. In this research, we develop a new heuristic on how to build BNs from sparse datasets in the context of its performance in a real-time recommendation system. This new heuristic is demonstrated using a market basket dataset and a real-time recommendation model where all items in the grocery store are RFID tagged and the carts are equipped with an RFID scanner. With this recommendation model, retailers are able to do real-time recommendations to customers based on the products placed in cart during a shopping event.

Two-Phase Collaborative Filtering Algorithm Based on Co-Clustering

This paper proposes a two-phase rating predicting framework that fuses co-clustering and non-negative matrix factorization method. First, it uses a novel co-clustering method (BlockClust) to divide the raw rating matrix into clusters much smaller than the original matrix. Then it employs weighted non-negative matrix factorization algorithm to predict the unknown ratings. In virtue of co-clustering preprocessing, this method achieves a higher predicting accuracy and efficiency on these low-dimensional and homogeneous sub-matrices. Moreover, it proposes three update schemes for the corresponding update scenarios in recommender systems. Finally, the proposed method is implemented together with seven types of related CF (collaborative filtering) methods. The comparisons show the efficiency of the proposed method and its potential in large real-time recommender systems.

실시간 컨텍스트 정보의 정량화 단계를 개선한 협력적 필터링

In general, recommendation systems quantify real-time context information obtained in the stage of collaborative filtering and use quantified context information in order to recommend services. But the recommendation systems can have problems of recommending inaccurate information because of lack of context information or classifying users into inaccurate groups because of simple classification works in the stage of quantification. In this paper, we solved the problems of lack of context information obtained in real-time by combining users` profile information used in the contents-based filtering and context information obtained in real-time. In addition, we tried collaborative filtering at the quantification stage by improving absolute classification methods to relative ones. As the result of experiments, this method improved prediction preference by 5.8% than real-time recommendation systems using context information in pure P2P environment.

Old instruments and devices

This paper presents a framework for estimating and updating user preferences in the context of app-based recommender systems. We specifically consider recommender systems which provide personalized menus of options to users. A Hierarchical Bayes procedure is applied in order to account for inter- and intra-consumer heterogeneity, representing random taste variations among individuals and among choice situations (menus) for a given individual, respectively. Three levels of preference parameters are estimated: population-level, individual-level and menu-specific. In the context of a recommender system, the estimation of these parameters is repeated periodically in an offline process in order to account for trends, such as changing market conditions. Furthermore, the individual-level parameters are updated in real-time as users make choices in order to incorporate the latest information from the users. This online update is computationally efficient which makes it feasible to embed it in a real-time recommender system. The estimated individual-level preferences are stored for each user and retrieved as inputs to a menu optimization model in order to provide recommendations. The proposed methodology is applied to both Monte-Carlo and real data. It is observed that the online update of the parameters is successful in improving the parameter estimates in real-time. This framework is relevant to various recommender systems that generate personalized recommendations ranging from transportation to e-commerce and online marketing, but is particularly useful when the attributes of the alternatives vary over time.