A two-stage location-sensitive and user preference-aware recommendation system

https://doi.org/10.1016/j.eswa.2021.116188Get rights and content

Highlights

  • Increasing the accuracy in estimating the quantitative/qualitative factors of service.

  • Services with the least response time and the highest level of confidentiality.

  • Supporting the expressed needs of cloud customers verbally and numerically.

Abstract

Nowadays, cloud customers use cloud services increasingly to satisfy their demands. Usually, a significant number of customers are immature and inexpert and cannot express their needs accurately and numerically. They usually express their needs verbally and in the form of linguistic terms. On the other hand, the experienced customers express their needs numerically and accurately. In this situation, a recommendation system can be considered as one of the most useful ideas to support all type of customers. However, current recommendation systems (e.g., collaborative filtering based recommendations) meet customer requests that are accurately and numerically expressed. To support all types of customers, the construction of a strong recommendation system to analysis the demands expressed by customers (experienced and inexperienced) and to recommend suitable services is vital. As another important matter, cloud customers and services have been geographically distributed. Identifying the location of customers and services has a significant effect on the quality of services offered to customers. Therefore, the recommendation system should consider the location of customers and services in order to provide better services. In this paper, we introduce an efficient method to construct a powerful recommendation system which can provide suitable services considering the preferences of the customer and their location. The proposed recommendation system comprises two algorithms. The first algorithm is a fuzzy clustering algorithm, named FCA, that can well classify the location of customers and services. The second algorithm is an iterative adaptive neural-fuzzy algorithm, named IANFRA, which receives the preferences of the customer along with their location and identifies suitable services based on the locations clustered by FCA and the demands of customers (experienced and inexperienced). Finally, the feasibility of the proposed method has validated in terms of accuracy and scalability through conducting extensive experiments on a real distributed service quality dataset WS-DREAM. The evaluation results illustrate that both the service recommendation accuracy in the prediction of quality of services and the scalability, when the volume of the dataset is huge, have been improved.

Introduction

A cloud recommendation system is a middleware which provides the desired cloud services for cloud customers using managing multiple cloud services and negotiation between the cloud providers and cloud consumers. In a cloud environment, the recommendation system manages uncertainty situation in cloud services, heterogeneity, and dimensionality. It can play a significant role in helping cloud customers to select the most appropriate cloud services with lower cost and more high quality (Chauhan et al., 2019, Elhabbash et al., 2019, Tordsson et al., 2012).

Based on pay-per-use logic in the cloud computing model, each user can use on-demand cloud services and just pay cost considering the amount of own consumption (Bhat et al., 2011, Weinman, 2016). In recent years, a drastic competitive between cloud providers have been constructed. Providers in different countries produce cloud services which are functionally equivalent. Consequently, predicting the value of quality of service (QoS) and selecting the most proper service among candidate services is a difficult task, since the quality of service highly depends on various factors such as the location of providers and users, the clarity in precise expressing service requirements, and etc. (Bushehrian and Mohammadi, 2016, Esposito et al., 2015a). Hitherto, a large number of methods for identifying and selecting the suitable cloud services have been presented (Ding et al., 2017, Esposito et al., 2015a, Sun et al., 2016, ur Rehman et al., 2011). However, the presented methods cannot support immature customers and cannot satisfy their demands.

It can be argued that the following factors in the cloud service recommendation field need special attention because these factors can significantly affect the quality of cloud services, the satisfaction of cloud customers, the reputation of providers, satisfying service level agreement (SLA) and so on.

  • The lack of specifying service requirements in precise terms: The use of cloud services is being grown by users to satisfy their requirements. Some cloud users are immature and inexperienced and cannot express their needs accurately and numerically. Hence, they use linguistic expressions (such as medium, high, low and so on) to express their needs. For example, the inexperienced users to use cloud infrastructure (IaaS) and to meet computing requirements may express their need as Hard Disk (HDD) = Medium, RAM = High, computing capacity = High, network bandwidth = Medium. Without a flexible recommendation system, such requests cannot be met.

  • Prevention of service level agreements (SLA) violation: In a dynamic cloud environment, contracts between customers and providers should be properly managed so that be prevented SLA violations. To arrive this goal, the proposed recommendation system tunes different fuzzy weights for each QoS so that leads to the prevention of SLA violation. Indeed, the proposed method uses a fuzzy value tuned by the neuro-fuzzy network instead utilization of a crisp value for a QoS metric.

  • Impacts of the geographical distribution of providers and customers on QoS: Some hidden factors such as the geographical location of cloud service customers or providers and the distance between them can easily affect QoS values. Hence, knowledge of the geographical location of customers and providers has a significant effect on recommendation system decision.

  • Differences in country laws regarding the use of cloud services: Since, the different providers and customers have geo-distribution, ensuring the privacy and security of cloud customer data is a major challenge, especially when data is located in other countries. Because different countries usually have different laws. Hence, guaranteeing confidentiality and security on data is difficult. Moreover, concerning legal borders in some situations, there is a restrict restriction to use cloud services out of country borders (El-Gazzar & Wahid, 2015). For example, considering the European Union (EU) decree, all personal data which are produced within the EU should not leave the EU. Unless the destination country has sufficient credibility to protect the privacy of the information (El-Gazzar and Wahid, 2015, Essén et al., 2018). Consequently, considering the location of providers and consumers and laws of different countries are essential to select cloud service. In this paper, the proposed method can find the closest services to any cloud customer and can protect the security and confidentiality of information a large extent. To address the mentioned issue, this paper proposes a new technique using an innovative neuro-fuzzy method which combines the advantages of the fuzzy logic with the generalizability and the prediction power of neural networks. The neuro-fuzzy method analyzes the specifications and requirements announced by the customer and can find the closest services to the customer and estimate their price.

In this paper, to more clarity, the proposed method has been presented to provide IaaS services so that the cloud customer can request own desired storage as two ways, numerical for example Hard Disk (HDD)= 13TB or as a linguistic term, for example, Hard Disk (HDD) = Medium. The fuzzy logic has been used to construct a recommendation system which can handle all types of customer requests i.e. numerical and linguistic terms. To better estimate the QoS and select a suitable service by the recommendation system, we merge fuzzy logic with neural network science. When a customer sends a cloud service request to a recommendation system, the recommendation system takes the service price and other quality characteristics from providers and considering the distance between the provider and customer and the preferences of customer such as the quality and price of services recommends a suitable service to the customer. In this method, after requesting a specific service by the customer, the recommendation system creates a sorted list of each feature of service according to the services announced by the providers. After that, the recommendation system defines a range of values for the features and divides the range of values into sections and denominates each section with a linguistic term such as high, low, medium.

Fig. 1 illustrates a fuzzy range for HDD instance which has been divided into three sections namely Low, Medium, and High. It is worth noting that, the recommendation system uses an iterative neuro-fuzzy learning algorithm to tune the fuzzy range of feature values during the decision making to select cloud services. In the last step of the decision-making process, the recommendation system converts its fuzzy decision to a crisp decision by applying the Defuzzification method. The function of the proposed method has been explained in Section 4 in more detail. The presented method can attain high accuracy in the selection of the most appropriate neighbor cloud services for customers so that, the response time be minimal and the offered cloud service has most adaptation with the demands of cloud customers. It is necessary to mention, the proposed method is suitable for all types of cloud customers (experienced and inexperienced users). Inexperienced customers can linguistically express their requirements and attain a proper service.

Hitherto, various techniques have been presented in the area of cloud service selection to identify and recommend the most suitable cloud service to customers (He et al., 2014, Jin et al., 2019, Luo et al., 2016a, Peng et al., 2018, Ren and Wang, 2018, Ryu et al., 2018a, Tang et al., 2016a, Yin et al., 2019, Yin et al., 2017, Zheng et al., 2009). However, existing cloud service selection methods have not considered the immature customers. In this paper, the mentioned challenge has been solved by combining the various powerful science i.e. the fuzzy logic science and neural-network science. In summary, the main contributions of paper are as follows:

  • A Location clustering technique has been proposed based on a fuzzy clustering algorithm (FCA) where the users can belong to two or more regions, and geographical locations are expressed in the form of latitude–longitude. A DoP (the degree of proximity) metric has been introduced where DoPij denotes the membership value of data point i (customer i) to cluster j (provider j). Using the FCA, the recommendation system can find providers close to the customers who are placed in a geolocation zone in the earth map.

  • An iterative adaptive neural fuzzy recommendation algorithm (IANFRA) has been proposed that can usefully manage the needs expressed verbally and numerically (a crisp value) and predict features and price of services based on the desired preferences of customers to recommend a suitable service.

  • Wide experiments have been conducted to examine the performance of the proposed method using two dataset. The first dataset is a large set of real web services called WS-DREAM (Zheng & Lyu, 2008) and the second dataset is an expanded dataset of cloud IaaS services which offered by https://www.digitalocean.com/ (2021).

The remainder of the paper is organized as follows. In Section 2, the basic concepts related to proposed method are introduced. In Section 3, the related works are reviewed and an overview of the proposed IANFRA is presented in Section 4. The experiments has been explained in Section 5 in detail and the evaluation of the presented recommendation system has been done in Section 6, and finally the conclusion is presented in Section 7.

Section snippets

Background

To describe the IANFRA, some basic definitions should be introduced for the concepts used in this article. These definitions cover two areas including the fuzzy clustering algorithm and the iterative adaptive neural fuzzy recommendation algorithm.

Related works

In this section, the related works in the area of cloud service selection and recommendation have been reviewed. Various studies have been conducted in the field of recommendation of cloud services to customers, nevertheless, the presented works have assumed that all cloud customers can express their own demands as exact and numerical. Therefore, an immature customer cannot receive proper services from these recommenders. According to the latest studies, existing approaches to provide a

The proposed recommendation system

As previously stated, in the field of cloud service selection and recommendation, most of the available studies presume all customers are expert and express their requirements, such as the amount of the hard disk (HDD) or RAM required for an IaaS service, in an explicit and precise numeric way (Chen and Ma, 2015, Chen et al., 2015, Jin et al., 2019, Kuang et al., 2018, Luo et al., 2016a, Ren and Wang, 2018, Ryu et al., 2018a, Shao et al., 2007, Tang et al., 2016a, Yin et al., 2019, Yin et al.,

Experiments

In this section, we have conducted comprehensive experiments to demonstrate the ability of the proposed recommendation system to identify and recommend appropriate services according to the location and preferences of users. All the experiments have been implemented in MATLAB R2019b on Windows10 64 bit OS with Intel Core i7-8750 2.20 GHz processor and 16 GB RAM. As mentioned stated, one of the primary advantages of the proposed recommender is satisfying the demands of all types of users

Evaluation

The purpose of evaluation is to answer the following questions:

(1) how is the performance of FCA compared with other popular clustering methods such as k-means algorithm and subtractive clustering algorithm? (2) How generalizable and scalable is the proposed method in large datasets? (3) What is the advantage of the proposed method compared with other related methods? and how is the performance of IANFRA compared with other methods?

Conclusion

In this paper a novel method has been proposed to identify and to recommend suitable cloud services considering the demands of the cloud users and the geolocation distribution of them. The proposed method uses a fuzzy clustering algorithm to find providers close to the users. Then, an iterative adaptive neural-fuzzy recommendation algorithm (IANFRA) is represented to identify and to recommend suitable services considering the preferences of the user. The immature users can easily express their

CRediT authorship contribution statement

Neda Mohammadi: Conceptualization, Investigation, Formal analysis, Validation, Visualization, Methodology, Writing – original draft, Writing. Abbas Rasoolzadegan: Conceptualization, Project administration, Supervision, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (86)

  • NilashiM. et al.

    A multi-criteria collaborative filtering recommender system for the tourism domain using expectation maximization (EM) and PCA–ANFIS

    Electronic Commerce Research and Applications

    (2015)
  • RadionovA. et al.

    Application of subtractive clustering for power transformer fault diagnostics

    Procedia Engineering

    (2015)
  • RenL. et al.

    An SVM-based collaborative filtering approach for top-n web services recommendation

    Future Generation Computer Systems

    (2018)
  • SuK. et al.

    TAP: A personalized trust-aware QoS prediction approach for web service recommendation

    Knowledge-Based Systems

    (2017)
  • SunL. et al.

    Cloud-FuSeR: Fuzzy ontology and MCDM based cloud service selection

    Future Generation Computer Systems

    (2016)
  • SyuY. et al.

    Time series forecasting for dynamic quality of web services: an empirical study

    Journal of Systems and Software

    (2017)
  • TordssonJ. et al.

    Cloud brokering mechanisms for optimized placement of virtual machines across multiple providers

    Future Generation Computer Systems

    (2012)
  • WuH. et al.

    Deviation-based neighborhood model for context-aware QoS prediction of cloud and IoT services

    Future Generation Computer Systems

    (2017)
  • WuH. et al.

    Collaborative QoS prediction with context-sensitive matrix factorization

    Future Generation Computer Systems

    (2018)
  • XuY. et al.

    Context-aware QoS prediction for web service recommendation and selection

    Expert Systems with Applications

    (2016)
  • ZhangY. et al.

    Exploring latent features for memory-based QoS prediction in cloud computing

  • ZhouQ. et al.

    Spatio-temporal context-aware collaborative QoS prediction

    Future Generation Computer Systems

    (2019)
  • AdomaviciusG. et al.

    Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions

    IEEE Transactions on Knowledge and Data Engineering

    (2005)
  • BelussoC.L. et al.

    Price modeling of iaas providers-an approach focused on enterprise application integration

  • BezdekJ.C. et al.

    FCM: The fuzzy c-means clustering algorithm computer & geosciences

    Volume

    (1984)
  • BhatM.A. et al.

    Cloud computing: A solution to geographical information systems(GIS)

    International Journal on Computer Science and Engineering

    (2011)
  • BushehrianO. et al.

    Association rule mining for failure recovery strategy selection in composite web services

    Journal of Telecommunication, Electronic and Computer Engineering (JTEC)

    (2016)
  • CaoJ. et al.

    Hybrid collaborative filtering algorithm for bidirectional web service recommendation

    Knowledge and Information Systems

    (2013)
  • ChenM. et al.

    A hybrid approach to web service recommendation based on qos-aware rating and ranking

    (2015)
  • ChenM. et al.

    A ranking-oriented hybrid approach to QoS-aware web service recommendation

  • ChenL. et al.

    Predicting quality of service via leveraging location information

    Complexity

    (2019)
  • ChunS.H. et al.

    Service models and pricing schemes for cloud computing

    Cluster computing

    (2014)
  • CzekalskiP.

    Evolution-fuzzy rule based system with parameterized consequences

    International Journal of Applied Mathematics and Computer Science

    (2006)
  • DastjerdiA.V. et al.

    Compatibility-aware cloud service composition under fuzzy preferences of users

    IEEE Transactions on Cloud Computing

    (2014)
  • DickersonK.B. et al.

    Music recommendation and query-by-content using self-organizing maps

  • El-Gazzar, R. F., & Wahid, F. (2015). Strategies for cloud computing adoption: insights from the Norwegian public...
  • ElangovanR. et al.

    Location-based social network recommendations with computational intelligence-based similarity computation and user check-in behavior

    Concurrency Computations: Practice and Experience

    (2020)
  • ElhabbashA. et al.

    Cloud brokerage: A systematic survey

    ACM Computing Surveys

    (2019)
  • EspositoC. et al.

    Smart cloud storage service selection based on fuzzy logic, theory of evidence and game theory

    IEEE Transactions on Computers

    (2015)
  • EspositoC. et al.

    Smart cloud storage service selection based on fuzzy logic, theory of evidence and game theory

    IEEE Transactions on Computers

    (2015)
  • FattahiH.

    Indirect estimation of deformation modulus of an in situ rock mass: an ANFIS model based on grid partitioning, fuzzy c-means clustering and subtractive clustering

    Geosciences Journal

    (2016)
  • FerrandezS.M. et al.

    Optimization of a truck-drone in tandem delivery network using k-means and genetic algorithm

    Journal of Industrial Engineering and Management (JIEM)

    (2016)
  • GhafouriS.H. et al.

    A survey on web service QoS prediction methods

    IEEE Transactions on Services Computing

    (2020)
  • Cited by (7)

    • An optimized recommendation framework exploiting textual review based opinion mining for generating pleasantly surprising, novel yet relevant recommendations

      2022, Pattern Recognition Letters
      Citation Excerpt :

      A user may lose interest in an RS if RS remains restricted to relevant or popular items and exclusively focuses on accuracy [4]. Instead, the RS should include a diverse, novel and unexpected set of preferences [5,6]. Diversity and novelty-based RS [7] avoid the overrated and similar recommendations.

    • News Recommendation with Multi-views Emotion Analysis

      2022, ACM International Conference Proceeding Series
    View all citing articles on Scopus

    The code (and data) in this article has been certified as Reproducible by Code Ocean: (https://codeocean.com/). More information on the Reproducibility Badge Initiative is available at https://www.elsevier.com/physical-sciences-and-engineering/computer-science/journals.

    View full text