highly accurate prediction algorithmDissertation_PPT.pptx

quality of services on web

1. Dr. Babasaheb Ambedkar Marathwada University, Aurangabad CSMSS CHH. SHAHU COLLEGE OF ENGINEERING, AURANGABAD DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING Student : Pallavi R. Kathar M.E (CSE) PRN No .:2014015201031031 A Presentation Of Dissertation – II Title “ A Highly Accurate Prediction Algorithm for Unknown Web Service QoS Values Under the Guidance of: Dr. S. P. Abhang Department of Computer Science & Engineering

2. Contents Introduction Existing Methodologies Proposed Work Algorithms Conclusion CSMSS, Chh. Shahu College of Engineering, Aurangabad

3. CSMSS, Chh . Shahu College of Engineering, Aurangabad Introduction Often we get confused between a website and a web service both of them are rendering some kind of service to the end user over the network but there exists a difference between the two and it is that, a website is meant for human consumption and a service which is meant for code consumption or application level consumption is called a web service. Web Services are software components designed to support interoperable machine to machine interaction over a network, and through standard Web protocol to provide services.

4. Quality of Service ( QOS ) is usually used to describe the non-functional characteristics of Web Services. With the increasing abundance of Web Services on the World Wide Web, studies on QoS become more and more attractive. Recommender systems employ Information Filtering technique that focuses on providing the recommendations of the items to the users that are likely to be of the user’s interest. CSMSS, Chh. Shahu College of Engineering, Aurangabad

5. Collaborative Filtering (CF) algorithms has been widely used to predict missing values in commercial recommender systems Collaborative filtering algorithms often require: 1. user’s active participation 2. An easy way to represent user’s interests to the system 3. Algorithms those are able to match people with similar interests CSMSS, Chh. Shahu College of Engineering, Aurangabad

6. Typically, the workflow of a collaborative filtering system is: A user expresses his or her preferences by rating items (e.g. books, movies or CDs) of the system. These ratings can be viewed as an approximate representation of the user's interest in the corresponding domain. The system matches this user’s ratings against other users’ and finds the people with most “similar” tastes. CSMSS, Chh. Shahu College of Engineering, Aurangabad

7. With similar users, the system recommends items that the similar users have rated highly but not yet being rated by this user (presumably the absence of rating is often considered as the unfamiliarity of an item). CSMSS, Chh. Shahu College of Engineering, Aurangabad

8. Existing Methodologies The QoS data of Web services are usually objective, meaning that existing collaborative filtering-based approaches are not always applicable for unknown QoS values. Most existing QoS prediction methods are inspired by these CF ideas, for which we call traditional CF methods to distinguish our proposed algorithm. CSMSS, Chh. Shahu College of Engineering, Aurangabad

9. CSMSS, Chh. Shahu College of Engineering, Aurangabad The existing CF based prediction methods for un-known QoS values have not realized the above differences between subjective and objective data and therefore cannot predict objective QoS values accurately. In allusion to this problem presents a highly accurate prediction algorithm (HAPA) for unknown Web service QoS values. HAPA is also CF-based, i.e. also use similar users and similar items to make prediction, but with fundamental changes from traditional CF approaches to adapt to the characteristics of objective QoS data.

10. Proposed work using Highly Accurate Prediction Algorithm (HAPA) which is Collaborative Filtering based algorithm, i.e., it actually includes user-based and item-based HAPAs. Both kind of HAPA can compose forecasts; though we constantly apply the grouping of two HAPAs to construct more correct forecasts. CSMSS, Chh. Shahu College of Engineering, Aurangabad

11. Pearson Correlation Coefficient(PCC) It is used between to measure user similarity in recommendation systems. It measures the similarity two service users based on the QoS values of Web services. Two service users have similar Web service usage experiences if the PCC value is positive and a negative PCC value indicates that their experiences are opposite. The value is null when two service users have no commonly invoked web service. CSMSS, Chh. Shahu College of Engineering, Aurangabad

12. CSMSS, Chh . Shahu College of Engineering, Aurangabad

13. CSMSS, Chh. Shahu College of Engineering, Aurangabad

14. CSMSS, Chh. Shahu College of Engineering, Aurangabad How Our Algorithm will work: Adaptive Matrix Factorization Algorithm: we propose our new QoS prediction approach, adaptive matrix factorization (AMF), which aims to be online, accurate, and scalable. To achieve this goal, our AMF approach integrates three techniques: data transformation, online learning, and adaptive weights.

15. The singular values are computed by a singular value decomposition (SVD) and then normalized so that the largest singular value is equal to 1. We can observe that except the first few largest singular values, most of them are close to 0. This observation indicates that both data matrices are approximately low-rank, which conforms to our low-rank assumption of matrix factorization. In our experiment, we set rank d = 10 CSMSS, Chh. Shahu College of Engineering, Aurangabad

16. Conclusion Through this project , We have utilized CF to predict unknown QoS values. Strictly speaking, our approach is essentially different from traditional CF which is not applicable to objective data prediction. Traditional CF is based on the premise that similar users have similar subjective experience on the same items, but this premise no longer applies for the objective data. Based on PCC characteristics, we proposed our HAPA. The prediction accuracy of HAPA was shown to outperform that of many of existing QOS prediction methods. CSMSS, Chh. Shahu College of Engineering, Aurangabad