International Journal on Artificial Intelligence Tools

In recent years, a new research area known as privacy preserving data mining (PPDM) has emerged and captured the attention of many researchers interested in preventing the privacy violations that may occur during data mining. In this paper, we provide a review of studies on PPDM in the context of association rules (PPARM). This paper systematically defines the scope of this survey and determines the PPARM models. The problems of each model are formally described, and we discuss the relevant approaches, techniques and algorithms that have been proposed in the literature. A profile of each model and the accompanying algorithms are provided with a comparison of the PPARM models.

The aim of the frequent subgraph mining task is to find frequently occurring subgraphs in a large graph database. However, this task is a thriving challenge, as graph and subgraph isomorphisms play a key role throughout the computations. Since subgraph isomorphism testing is a hard problem, subgraph miners are exponential in runtime. To alleviate the complexity issue, we propose to introduce a bias in the projection operator and instead of using the costly subgraph isomorphism projection, one can use a polynomial projection having a semantically-valid structural interpretation. This paper presents a new projection operator for graphs named AC-projection, which exhibits nice theoretical complexity properties. We study the size of the search space as well as some practical properties of the projection operator. We also introduce a novel breadth-first algorithm for frequent AC-reduced subgraphs mining. Then, we prove experimentally that we can achieve an important performance gain (polynomial complexity projection) without or with non-significant loss of discovered patterns in terms of quality.

In this study, a hybrid of Quantum Evolutionary and Artificial Immune Algorithms (QIA) is proposed for solving Multiobjective Flexible Job Shop Scheduling Problem (MFJSSP). This problem is formulated as three-objective problem which minimizes completion time (makespan), critical machine workload and total work load of all machines. The quantum coding is shown to improve the immune strategy. The proposed algorithm overcomes the problem by increasing the speed of convergence and diversity of population. Three benchmarks of Kacem and Brandimart are examined to evaluate the performance of the proposed algorithm. The experimental results show a better performance in comparison to other approaches.

This paper presents a privacy preserving protocol for the computation of a Radial Basis Function (RBF) neural network model between N participants which share horizontally partitioned datasets. The RBF model is used for regression analysis tasks. The novel aspect of the proposed protocol lies to the fact that it assumes a malicious user model and does not use homomorphic cryptographic methods, which are inherently only suited for a semi-trusted user environment. The performance analysis shows that the communication overhead is low enough to warranty its use while the computational complexity is identical in most cases with the centralized computation scenario (e.g. a trusted third party). The accuracy of the output model is only marginally subpar to a centralized computation on the union of all datasets.

For an industrial system reliability, availability and maintainability (RAM) analysis play an important role in any design modification for achieving its optimum performance. However, it is difficult to predict these parameters by using available and imprecise data up to a desired degree of accuracy. For this, a novel technique named as an artificial bee colony based Lambda-Tau has been presented for computing these parameters by utilizing available or collected data up to a desired degree of accuracy. In this technique expression of RAM parameters are calculated by Lambda-Tau methodology and their corresponding membership functions are computed by formulating a nonlinear programming problem. A generalized RAM-Index has been used for ranking the components of the system based on its performance for improving the system productivity. The presented approach has been investigated through a case study of washing unit of paper industry and computed results are compared with existing Lambda-Tau and evolutionary algorithm techniques.

At present many algorithms for mining association rules have been proposed, but most of them are only suitable for discovering specific frequent itemsets from characteristic data sets on the appointed environments, namely, these algorithms are not general enough when mining association rules. In this paper, a general framework based on composite granules for mining association rules is proposed, which is a general data mining model without appointed restriction from frequent itemsets, data sets or mining environments and so on. An iterative method is efficiently applied to the general mining framework for discovering frequent itemsets, which adopts repartitioning frequent attributes to iteratively reconstruct the mixed radix information system for reducing a relational database. In order that the framework for discovering frequent itemsets has a generality, in discussing and establishing the general mining framework, this paper introduces a novel conception and data model, namely, a mixed radix information system is applied to describe a relational database, and a composite granules is used to build a specific relationship between an information system and a mixed radix information system, which can hold the same extension and simultaneously exist in two different information systems. The mixed radix information system can help the general framework to reduce information data and improve the performance of the framework for generating frequent itemsets. The composite granules model can create a relationship between an information granule and a digital information granule, and help the framework for computing the support to avoid reading the database repeatedly or using the complex data structure. Finally, a new taxonomy is presented to verify the generality and the high efficiency of the mining framework and all the experiments based on the taxonomy indicate that the general mining framework has the required generality, and the performance of the framework is better than these classical mining frameworks.

Parkinson's disease is a chronic, progressive, age-related, neurodegenerative disorder that affects a large population around the world. A mathematical model for Parkinson's disease is presented using Fuzzy Cognitive Maps (FCMs). Basic theories of FCMs are reviewed and presented. Decision Support Systems (DSS) for medical problems are reviewed. Non-linear Hebbian learning techniques are considered in studying Medical problems and a generic algorithm is presented. The proposed method used the knowledge of a number of experts and simulations were performed obtaining interesting results. Comparisons of the results of the proposed method, both by making use and not making use of learning algorithms, are presented. Some interesting future research directions are mentioned.

Osteoporosis is a disease of bones that leads to an increased risk of fracture and it is characterized by low bone mineral density and micro-architectural deterioration of bone tissue. In this article, the dataset consists of 3426 subjects (1083 pathological and 2343 healthy cases) whose diagnosis was based on laboratory and osteal bone densitometry examination. In all cases, four diagnostic factors for osteoporosis risk prediction, namely age, sex, height and weight were stored for later evaluation with the selected classifiers. In order to categorize subjects into two classes (osteoporosis, nonosteoporosis), twenty machine learning techniques were assessed, based on their popularity and frequency in biomedical engineering problems. All classifiers have been evaluated using the wellknown 10-fold cross validation method and the results are reported analytically. In addition, a feature selection method identified that with the use of only two diagnostic factors (age and weight), similar performance could be achieved. The scope of the proposed exhaustive methodology is to assist therapists in osteoporosis prediction, avoiding unnecessary further testing with bone densitometry.

Geographical referencing of data and resources on the Web has become prevalent. Discovering and linking this information poses eminent research challenges to the geospatial semantic web, with regards to the representation and manipulation of information on geographic places. Towards addressing these challenges, this work explores the potential of the current semantic web languages and tools. In particular, an integrated logical framework of rules and ontologies, using current W3C standards, is assessed for modeling geospatial ontologies of place and for encoding both symbolic and geometric references to place locations. Spatial reasoning is incorporated in the framework to facilitate the deduction of implicit spatial relations and for expressing spatial integrity constraints. The logical framework is extended with geo-computation engines that offer more effective manipulation of geometric information. Example data sets mined from web resources are used to demonstrate and evaluate the framework, offering insights to its potentials and limitations.