Mojtaba Salehi

Hybrid recommendation approach for learning material based on sequential pattern of the accessed material and the learner’s preference tree

Abstract The explosion of the learning materials in personal learning environments has caused difficulties to locate appropriate learning materials to learners. Personalized recommendations have been used to support the activities of learners in personal learning environments and this technology can deliver suitable learning materials to learners. In order to improve the quality of recommendations, this research considers the multidimensional attributes of material, rating of learners, and the order and sequential patterns of the learner’s accessed material in a unified model. The proposed approach has two modules. In the sequential-based recommendation module, latent patterns of accessing materials are discovered and presented in two formats including the weighted association rules and the compact tree structure (called Pattern-tree). In the attribute-based module, after clustering the learners using latent patterns by K-means algorithm, the learner preference tree (LPT) is introduced to consider the multidimensional attributes of materials, rating of learners, and also order of the accessed materials. The mixed, weighted, and cascade hybrid methods are employed to generate the final combined recommendations. The experiments show that the proposed approach outperforms the previous algorithms in terms of precision, recall, and intra-list similarity measure. The main contributions are improvement of the recommendations’ quality and alleviation of the sparsity problem by combining the contextual information, including order and sequential patterns of the accessed material, rating of learners, and the multidimensional attributes of materials.

Personalized recommendation of learning material using sequential pattern mining and attribute based collaborative filtering

Material recommender system is a significant part of e-learning systems for personalization and recommendation of appropriate materials to learners. However, in the existing recommendation algorithms, dynamic interests and multi-preference of learners and multidimensional-attribute of materials are not fully considered simultaneously. Moreover, these algorithms cannot effectively use the learner’s historical sequential patterns of material accessing in recommendation. For addressing these problems and improving the accuracy and quality of recommendation, a new material recommender system framework based on sequential pattern mining and multidimensional attribute-based collaborative filtering (CF) is proposed. In the sequential pattern based approach, modified Apriori and PrefixSpan algorithms are implemented to discover latent patterns in accessing of materials and use them for recommendation. Leaner Preference Tree (LPT) is introduced to take into account multidimensional-attribute of materials, and learners’ rating and model dynamic and multi-preference of learners in the multidimensional attribute-based CF approach. Finally, the recommendation results of two approaches are combined using cascade, weighted and mixed methods. The proposed method outperforms the previous algorithms on the classification accuracy measures and the learner’s real learning preference can be satisfied accurately according to the real-time up dated contextual information.

On-line analysis of out-of-control signals in multivariate manufacturing processes using a hybrid learning-based model

Advanced automatic data acquisition is now widely adopted in manufacturing industries and it is common to monitor several correlated quality variables simultaneously. Most of multivariate quality control charts are effective in detecting out-of-control signals based upon an overall statistics in multivariate manufacturing processes. The main problem of such charts is that they can detect an out-of-control event but do not directly determine which variable or group of variables has caused the out-of-control signal and what is the magnitude of out of control. This study presents a hybrid learning-based model for on-line analysis of out-of-control signals in multivariate manufacturing processes. This model consists of two modules. In the first module using a support vector machine-classifier, type of unnatural pattern can be recognized. Then by using three neural networks for shift mean, trend and cycle it can be recognized magnitude of mean shift, slope of trend and cycle amplitude for each variable simultaneously in the second module. The performance of the proposed approach has been evaluated using two examples. The output generated by trained hybrid model is strongly correlated with the corresponding actual target value for each quality characteristic. The main contributions of this work are recognizing the type of unnatural pattern and classification major parameters for shift, trend and cycle and for each variable simultaneously by proposed hybrid model.

On line detection of mean and variance shift using neural networks and support vector machine in multivariate processes

The effective recognition of unnatural control chart patterns (CCPs) is one of the most important tools to identify process problems. In multivariate process control, the main problem of multivariate quality control charts is that they can detect an out of control event but do not directly determine which variable or group of variables has caused the out of control signal and how much is the magnitude of out of control. Recently machine learning techniques, such as artificial neural networks (ANNs), have been widely used in the research field of CCP recognition. This study presents a modular model for on-line analysis of out of control signals in multivariate processes. This model consists of two modules. In the first module using a support vector machine (SVM)-classifier, mean shift and variance shift can be recognized. Then in the second module, using two special neural networks for mean and variance, it can be recognized magnitude of shift for each variable simultaneously. Through evaluation and comparison, our research results show that the proposed modular performs substantially better than the traditional corresponding control charts. The main contributions of this work are recognizing the type of unnatural pattern and classifying the magnitude of shift for mean and variance in each variable simultaneously.

Attribute-based recommender system for learning resource by learner preference tree

In recent years, with growth of online learning technology, a huge amount of e-learning resources have been generated in various media formats. This growth has caused difficulty of locating appropriate learning resources to learners. A personalized recommendation is an enabling mechanism to overcome information overload occurred in the new learning environments and deliver suitable learner resources to learners. Since users express their opinions based on some specific attributes of items, this paper considers contextual information including attributes of learning resources and rating of learner simultaneously to address some problem such as sparsity and cold start problem and also improve the quality on recommendations. Learning Tree (LT) is introduced that can model the interest of learners based on attributes of learning resources in multidimensional space using learner historical accessed resources. Then, using a new similarity measure between learners, recommendations are generated. The experimental results show that our proposed method outperforms current algorithms and alleviates problems such as cold-start and sparsity.

Attribute-based collaborative filtering using genetic algorithm and weighted C-means algorithm

Recommender system technology can assist customers of a company to choose an appropriate product or service after learning their preferences. But this technology suffers from some problems such as scalability and sparsity. Since users express their opinions implicitly based on some specific attributes of items, this paper proposes a collaborative filtering algorithm based on attributes of items to address these problems. Attributes weight vector for each user is considered as a chromosome in genetic algorithm. This algorithm optimises the weights according to historical rating. A weighted C-means algorithm also is introduced to cluster users based on the optimised attributes weight vector. Finally, recommendation is generated by a user based similarity in each cluster. The experimental results show that our proposed method outperforms current algorithms and can perform superiorly and alleviates problems such as sparsity and precision quality. The main contribution of this paper is addressing sparsity problem using attribute weighting and scalability problem using weighted C-means algorithm.

A hybrid recommendation approach based on attributes of products using genetic algorithm and naive Bayes classifier

Recommender system technology can present personalised offers to customers of companies. This technology suffers from the cold-start and sparsity problems. On the other hand, in most researches, less attention has been paid to users preferences varieties in different product categories and also explicit and implicit attributes of products. Since users express their opinions implicitly based on some specific attributes of products, this paper proposes a hybrid recommendation approach based on attributes of products to address these problems. After product category and taxonomy formation and attributes extraction for each category, explicit-based module provides recommendations through naive Bayes classifier. Implicit-based module considers the weight vector of implicit attributes for users as chromosomes in genetic algorithm. This algorithm optimises the weights according to historical rating. Finally, recommendations are generated using the results of two modules. The main contributions are addressing sparsity and cold-start problem using naive Bayes classifier and weight optimisation by genetic algorithm.