Generating Highly Accurate Predictions for Missing QoS Data via Aggregating Nonnegative Latent Factor Models

Abstract

Automatic Web-service selection is an important research topic in the domain of service computing. During this process, reliable predictions for quality of service (QoS) based on historical service invocations are vital to users. This work aims at making highly accurate predictions for missing QoS data via building an ensemble of nonnegative latent factor (NLF) models. Its motivations are: 1) the fulfillment of nonnegativity constraints can better represent the positive value nature of QoS data, thereby boosting the prediction accuracy and 2) since QoS prediction is a learning task, it is promising to further improve the prediction accuracy with a carefully designed ensemble model. To achieve this, we first implement an NLF model for QoS prediction. This model is then diversified through feature sampling and randomness injection to form a diversified NLF model, based on which an ensemble is built. Comparison results between the proposed ensemble and several widely employed and state-of-the-art QoS predictors on two large, real data sets demonstrate that the former can outperform the latter well in terms of prediction accuracy.

An Effective Scheme for QoS Estimation via Alternating Direction Method-Based Matrix Factorization

“…For an MF-based QoS estimator, the fundamental data source is a user-service QoS-matrix as defined in [4][5][6][7][8][9][10][11]20]: Definition 1. Given a user set U and a service set S, a user-service QoS-matrix Q is a |U|×|S| matrix consisting of the historical records of a specified QoS-property by U on S where each known entry q u,s denotes the QoS-record by user u on service s. As mentioned before, Q is highly incomplete due to the impossibility for a user to invoke all services from S. Let Q K and Q U denote the known and unknwon entry sets of Q respectively, we have the following problem [4][5][6][7][8][9][10][11]20]: Definition 2.…”

“…Given a user set U and a service set S, a user-service QoS-matrix Q is a |U|×|S| matrix consisting of the historical records of a specified QoS-property by U on S where each known entry q u,s denotes the QoS-record by user u on service s. As mentioned before, Q is highly incomplete due to the impossibility for a user to invoke all services from S. Let Q K and Q U denote the known and unknwon entry sets of Q respectively, we have the following problem [4][5][6][7][8][9][10][11]20]: Definition 2. Given Q, the problem of MF-based QoS estimation is to build a rank-f approximation Q to Q based on Q K , such that the most accurate estimate ,u s q of each unknown entry q u,s ∈Q U is generated.…”

“…Another widely-adopted kind of approaches to CFbased QoS estimation is via matrix factorization (MF) [16][17][18][19][20][21][22]. A QoS estimator of this kind generally works by a) modeling the historical QoS data on a given QoS metric, i.e., request time, with a user-service matrix.…”

“…As unveiled by recent research in the recommender system community [12][13][14][15][16][17][18][19], MFbased CF models are highly accurate and scalable for many cases. They are also very effective in addressing the problem of QoS estimation [8][9][10][11]20]. However, motivated by practical requirements, it is still desired to further ---------------- I improve the accuracy of MF-based QoS estimators.…”

“…Nonetheless, it cost much memory to store the pair-wise factors for identifying the neighbors of each user/service in real applications. Luo et al [19,20] propose to ensemble a set of diversified non-negative latent factor models to achieve a highly accurate QoS estimator. This ensemble model is able to achieve high prediction accuracy.…”