Kleinberg introduced an axiomatic system for clustering functions. Out of three axioms, he proposed, two (scale invariance and consistency) are concerned with data transformations that should produce the same clustering under the same clustering function. The so-called consistency axiom provides the broadest range of transformations of the data set. Kleinberg claims that one of the most popular clustering algorithms, k-means does not have the property of consistency. We challenge this claim by pointing at invalid assumptions of his proof (infinite dimensionality) and show that in one dimension in Euclidean space the k-means algorithm has the consistency property. We also prove that in higher dimensional space, k-means is, in fact, inconsistent. This result is of practical importance when choosing testbeds for implementation of clustering algorithms while it tells under which circumstances clustering after consistency transformation shall return the same clusters. Two types of remedy are proposed: gravitational consistency property and dataset consistency property which both hold for k-means and hence are suitable when developing the mentioned testbeds.

We propose two new topic modeling methods for sequential documents based on hybrid inter-document topic dependency. Topic modeling for sequential documents is the basis of many attractive applications such as emerging topic clustering and novel topic detection. For these tasks, most of the existing models introduce inter-document dependencies between topic distributions. However, in a real situation, adjacent emerging topics are often intertwined and mixed with outliers. These single-dependency based models have difficulties in handling the topic evolution in such multi-topic and outlier mixed sequential documents. To solve this problem, our first method considers three kinds of topic dependencies for each document to handle its probabilities of belonging to a fading topic, an emerging topic, or an independent topic. Secondly, we extend our first method by considering fine-grained dependencies in a given context for more complex topic evolution sequences. Our experiments conducted on six standard datasets on topic modeling show that our proposals outperform state-of-the-art models in terms of the accuracy of topic modeling, the quality of topic clustering, and the effectiveness of outlier detection.

Nowadays, a vast amount of clinical data scattered across different sites on the Internet hinders users from finding helpful information for their well-being improvement. Besides, the overload of medical information (e.g., on drugs, medical tests, and treatment suggestions) have brought many difficulties to medical professionals in making patient-oriented decisions. These issues raise the need to apply recommender systems in the healthcare domain to help both, end-users and medical professionals, make more efficient and accurate health-related decisions. In this article, we provide a systematic overview of existing research on healthcare recommender systems. Different from existing related overview papers, our article provides insights into recommendation scenarios and recommendation approaches. Examples thereof are food recommendation, drug recommendation, health status prediction, healthcare service recommendation, and healthcare professional recommendation. Additionally, we develop working examples to give a deep understanding of recommendation algorithms. Finally, we discuss challenges concerning the development of healthcare recommender systems in the future.

Explainable recommendations have drawn more attention from both academia and industry recently, because they can help users better understand recommendations (i.e., why some particular items are recommended), therefore improving the persuasiveness of the recommender system and users’ satisfaction. However, little work has been done to provide explanations from the angle of a user’s contextual situations (e.g., companion, season, and destination if the recommendation is a hotel). To fill this research gap, we propose a new context-aware recommendation algorithm based on supervised attention mechanism (CAESAR), which particularly matches latent features to explicit contextual features as mined from user-generated reviews for producing context-aware explanations. Experimental results on two large datasets in hotel and restaurant service domains demonstrate that our model improves recommendation performance against the state-of-the-art methods and furthermore is able to return feature-level explanations that can adapt to the target user’s current contexts.

Employee churn (ECn) is a crucial problem for any organization that adversely affects its overall revenue and brand image. Many machine learning (ML) based systems have been developed to solve the ECn problem. However, they miss out on some essential issues such as employee categorization, category-wise churn prediction, and retention policy for effectively addressing the ECn problem. By considering all these issues, we propose, in this paper, a multi-attribute decision making (MADM) based scheme coupled with ML algorithms. The proposed scheme is referred as employee churn prediction and retention (ECPR). We first design an accomplishment-based employee importance model (AEIM) that utilizes a two-stage MADM approach for grouping the employees in various categories. Preliminarily, we formulate an improved version of the entropy weight method (IEWM) for assigning relative weights to the employee accomplishments. Then, we utilize the technique for order preference by similarity to ideal solution (TOPSIS) for quantifying the importance of the employees to perform their class-based categorization. The CatBoost algorithm is then applied for predicting class-wise employee churn. Finally, we propose a retention policy based on the prediction results and ranking of the features. The proposed ECPR scheme is tested on a benchmark dataset of the human resource information system (HRIS), and the results are compared with other ML algorithms using various performance metrics. We show that the system using the CatBoost algorithm outperforms other ML algorithms.

Knowledge graphs have become a popular formalism for representing entities and their properties using a graph data model, e.g., the Resource Description Framework (RDF). An RDF graph comprises entities of the same type connected to objects or other entities using labeled edges annotated with properties. RDF graphs usually contain entities that share the same objects in a certain group of properties, i.e., they match star patterns composed of these properties and objects. In case the number of these entities or properties in these star patterns is large, the size of the RDF graph and query processing are negatively impacted; we refer these star patterns as frequent star patterns. We address the problem of identifying frequent star patterns in RDF graphs and devise the concept of factorized RDF graphs, which denote compact representations of RDF graphs where the number of frequent star patterns is minimized. We also develop computational methods to identify frequent star patterns and generate a factorized RDF graph, where compact RDF molecules replace frequent star patterns. A compact RDF molecule of a frequent star pattern denotes an RDF subgraph that instantiates the corresponding star pattern. Instead of having all the entities matching the original frequent star pattern, a surrogate entity is added and related to the properties of the frequent star pattern; it is linked to the entities that originally match the frequent star pattern. Since the edges between the entities and the objects in the frequent star pattern are replaced by edges between these entities and the surrogate entity of the compact RDF molecule, the size of the RDF graph is reduced. We evaluate the performance of our factorization techniques on several RDF graph benchmarks and compare with a baseline built on top gSpan, a state-of-the-art algorithm to detect frequent patterns. The outcomes evidence the efficiency of proposed approach and show that our techniques are able to reduce execution time of the baseline approach in at least three orders of magnitude. Additionally, RDF graph size can be reduced by up to 66.56% while data represented in the original RDF graph is preserved.

Query Expansion (QE) is widely applied to improve the retrieval performance of ad-hoc search, using different techniques and several data sources to find expansion terms. In Information Retrieval literature, selecting expansion terms remains a challenging task that relies on the extraction of term relationships. In this paper, we propose a new learning to rank-based query expansion model. The main idea behind is that, given a query and the set of its related ARs, our model ranks these ARs according to their relevance score regarding to this query and then selects the most suitable ones to be used in the QE process. Experiments are conducted on three test collections, namely: CLEF2003, TREC-Robust and TREC-Microblog, including long, hard and short queries. Results showed that the retrieval performance can be significantly improved when the ARs ranking method is used compared to other state of the art expansion models, especially for hard and long queries.

Question Answering (QA) systems aim at supplying precise answers to questions, posed by users in a natural language form. They are used in a wide range of application areas, from bio-medicine to tourism. Their underlying knowledge source can be structured data (e.g. RDF graphs and SQL databases), unstructured data in the form of plain text (e.g. textual excerpts from Wikipedia), or combinations of the above. In this paper we survey the recent work that has been done in the area of stateless QA systems with emphasis on methods that have been applied in RDF and Linked Data, documents, and mixtures of these. We identify the main challenges, we categorize the existing approaches according to various aspects, we review 21 recent systems, and 23 evaluation and training datasets that are most commonly used in the literature categorized according to the type of the domain, the underlying knowledge source, the provided tasks, and the associated evaluation metrics.