Traffic forecasting has attracted widespread attention recently. In reality, traffic data usually contains missing values due to sensor or communication errors. The Spatio-temporal feature in traffic data brings more challenges for processing such missing values, for which the classic techniques (e.g., data imputations) are limited: (1) in temporal axis, the values can be randomly or consecutively missing; (2) in spatial axis, the missing values can happen on one single sensor or on multiple sensors simultaneously. Recent models powered by Graph Neural Networks achieved satisfying performance on traffic forecasting tasks. However, few of them are applicable to such a complex missing-value context. To this end, we propose GCN-M, a Graph Convolutional Network model with the ability to handle the complex missing values in the Spatio-temporal context. Particularly, we jointly model the missing value processing and traffic forecasting tasks, considering both local Spatio-temporal features and global historical patterns in an attention-based memory network. We propose as well a dynamic graph learning module based on the learned local-global features. The experimental results on real-life datasets show the reliability of our proposed method.

The term data quality refers to measuring the fitness of data regarding the intended usage. Poor data quality leads to inadequate, inconsistent, and erroneous decisions that could escalate the computational cost, cause a decline in profits, and cause customer churn. Thus, data quality is crucial for researchers and industry practitioners.Different factors drive the assessment of data quality. Data context is deemed one of the key factors due to the contextual diversity of real-world use cases of various entities such as people and organizations. Data used in a specific context (e.g., an organization policy) may need to be more efficacious for another context. Hence, implementing a data quality assessment solution in different contexts is challenging.Traditional technologies for data quality assessment reached the pinnacle of maturity. Existing solutions can solve most of the quality issues. The data context in these solutions is defined as validation rules applied within the ETL (extract, transform, load) process, i.e., the data warehousing process. In contrast to traditional data quality management, it is impossible to specify all the data semantics beforehand for big data. We need context-aware data quality rules to detect semantic errors in a massive amount of heterogeneous data generated at high speed. While many researchers tackle the quality issues of big data, they define the data context from a specific standpoint. Although data quality is a longstanding research issue in academia and industries, it remains an open issue, especially with the advent of big data, which has fostered the challenge of data quality assessment more than ever.This article provides a scoping review to study the existing context-aware data quality assessment solutions, starting with the existing big data quality solutions in general and then covering context-aware solutions. The strength and weaknesses of such solutions are outlined and discussed. The survey showed that none of the existing data quality assessment solutions could guarantee context awareness with the ability to handle big data. Notably, each solution dealt only with a partial view of the context. We compared the existing quality models and solutions to reach a comprehensive view covering the aspects of context awareness when assessing data quality. This led us to a set of recommendations framed in a methodological framework shaping the design and implementation of any context-aware data quality service for big data. Open challenges are then identified and discussed.

International audienceDue to the undeniable advantage of prediction and proactivity, many research areas and industrial applications are accelerating the pace to keep up with data science and predictive analytics. However and due to three well-known facts, the reactive Complex Event Processing (CEP) technology might lag behind when prediction becomes a requirement. 1st fact: The one and only inference mechanism in this domain is totally guided by CEP rules. 2nd fact: The only way to define a CEP rule is by writing it manually with the help of a human expert. 3rd fact: Experts tend to write reactive CEP rules, because and regardless of the level of expertise, it is nearly impossible to manually write predictive CEP rules. Combining these facts together, the CEP is---and will stay--- a reactive computing technique. Therefore in this article, we present a novel data mining-based approach that automatically learns predictive CEP rules. The approach proposes a new learning algorithm where complex patterns from multivariate time series are learned. Then at run-time, a seamless transformation into the CEP world takes place. The result is a ready-to-use CEP engine with enrolled predictive CEP rules. Many experiments on publicly-available data sets demonstrate the effectiveness of our approach

In the big data domain, data quality assessment operations are often complex and must be implementable in a distributed and timely manner. This article tries to generalize the quality assessment operations by providing a new ISO-based declarative data quality assessment framework (BIGQA). BIGQA is a flexible solution that supports data quality assessment in different domains and contexts. It facilitates the planning and execution of big data quality assessment operations for data domain experts and data management specialists at any phase in the data life cycle. This work implements BIGQA to demonstrate its ability to produce customized data quality reports while running efficiently on parallel or distributed computing frameworks. BIGQA generates data quality assessment plans using straightforward operators designed to handle big data and guarantee a high degree of parallelism when executed. Moreover, it allows incremental data quality assessment to avoid reading the whole dataset each time the quality assessment operation is required. The result was validated using radiation wireless sensor data and Stack Overflow users’ data to show that it can be implemented within different contexts. The experiments show a 71% performance improvement over a 1 GB flat file on a single processing machine compared with a non-parallel application and a 75% performance improvement over a 25 GB flat file within a distributed environment compared to a non-distributed application.

Recent research initiatives in the domain of business process management such as process intelligence, monitoring, and mining have shown significant results in automated process environments. However, such techniques fall short to provide efficient solutions and support for non-fully automated business processes i.e., processes embodying dynamic, continuous, and manual activities such as in logistics. More precisely, things turn to be very challenging when it comes to consider the monitoring and violation predictions through-out context-dependent and manual processes. Unlike current initiatives that mainly focus on the model of the process as a whole, we shift in this work towards instance-based and specific processing for each activity. We showcase a contextualized template-driven framework called the buttery, along side its architecture that could address the needs of continuous monitoring and prediction. Satisfactory results from evaluations on real data demonstrate the effectiveness of our framework.