The goal of service selection is to select services that satisfy user’s requirements from candidate services with the same function and different qualities of service (QoS). Traditional service selection methods require users to provide the weight for each QoS attribute, but users sometimes cannot provide accurate weights in practice. Moreover, because service providers may be untrustworthy, they may not provide reliable QoS values. Under these situations, the traditional service selection methods will be not very effective. To address this problem, we propose a service selection method based on ordinal classification for historical records. In this method, both QoS attributes and user-given ratings in historical records are considered as ordinal values, and an ordinal classification model will be learned from these data. For the historical records without user-given ratings, their ratings can be predicted by this model. Finally the services will be selected based on these ratings. The proposed method can work well even without the QoS attribute weights provided by user and the QoS values provided by service provider. We compare the proposed method with three weighted-based service selection methods and eight classification-based methods, which demonstrates that the proposed method can obtain better service selection results and have the best robustness.

With increasing adoption and presence of Web services, service composition becomes an effective way to construct software applications. Composite services need to satisfy both the functional and the non-functional requirements. Traditional methods usually assume that the quality of service (QoS) and the behaviors of services are deterministic, and they execute the composite service after all the component services are selected. It is difficult to guarantee the satisfaction of user constraints and the successful execution of the composite service. This paper models the constraint-satisfied service composition (CSSC) problem as a Markov decision process (MDP), namely CSSC-MDP, and designs a Q-learning algorithm to solve the model. CSSC-MDP takes the uncertainty of QoS and service behavior into account, and selects a component service after the execution of previous services. Thus, CSSC-MDP can select the globally optimal service based on the constraints which need the following services to satisfy. In the case of selected service failure, CSSC-MDP can timely provide the optimal alternative service. Simulation experiments show that the proposed method can successfully solve the CSSC problem of different sizes. Comparing with three representative methods, CSSC-MDP has obvious advantages, especially in terms of the success rate of service composition.

Monotonic classification is a kind of classification task in which a monotonicity constraint exist between features and class, i.e., if sample xihas a higher value in each feature than sample xj, it should be assigned to a class with a higher level than the level of xj's class. Several methods have been proposed, but they have some limits such as with limited kind of data or limited classification accuracy. In our former work, the classification accuracy on monotonic classification has been improved by fusing monotonic decision trees, but it always has a complex classification model. This work aims to find a monotonic classifier to process both nominal and numeric data by fusing complete monotonic decision trees. Through finding the completed feature subsets based on discernibility matrix on ordinal dataset, a set of monotonic decision trees can be obtained directly and automatically, on which the rank is still preserved. Fewer decision trees are needed, which will serve as base classifiers to construct a decision forest fused complete monotonic decision trees. The experiment results on 10 datasets demonstrate that the proposed method can reduce the number of base classifiers effectively and then simplify classification model, and obtain good classification performance simultaneously.