Companies often specify the intended behaviour of their business processes in a process model. Conformance checking techniques allow us to assess what degree such models and corresponding execution data correspond one another. In recent years, alignments have proven extremely useful for calculating conformance statistics. Existing compute been developed be used an offline, posteriori setting. However, we are interested observing deviations at moment they occur, rather than days, weeks or...

Measuring the quality of a clustering algorithm has shown to be as important itself. It is crucial part choosing that performs best for an input data. Streaming data have many features make them much more challenging than static ones. They are endless, varying and emerging with high speeds. This raised new challenges algorithms well their evaluation measures. Up till now, external measures were exclusively used validating stream algorithms. While validation requires ground truth which not...

Customer journey analysis is a hot topic in marketing. Understanding how the customers behave crucial and considered as one of key drivers business success. To best our knowledge, data-driven approach to analyze customer still missing. For instance, web analytics tools like Google Analytics provide an oversimplified version user behavior, focusing more on frequency page visits rather than discovering process visit itself. On other hand, maps have shown their usefulness, but they need be...

Process mining is an emerging research area that applies the well-established data solutions to challenging business process modeling problems. Mining streams of processes in real time as they are generated a necessity obtain instant knowledge from big data. In this paper, we introduce efficient approach for exploring and counting fragments stream events infer model using Heuristics Miner algorithm. Our novel approach, called Str ProM, builds prefix-trees extract sequential patterns stream....

Process mining is an emerging data task of gathering valuable knowledge out the huge collections business operation data.Despite its relatively young age, it has successfully provided many new insights into workflows using established techniques.Recently, with improvements in technologies sensoring, collection and storing data, a big demand for both shorter times adaptive models streaming process events arose.This initiated field stream very recently.Drifts underlying concepts processes are...

Process mining focuses on applying data techniques over business process data. Recently, with the improvements in sensoring, collection, and storage of event data, a big demand for both shorter time adaptive models streaming events arose. This increased interest mining. Some within this field attempt to identify drifts (change points) from evolving streams. Existing work supervised unsupervised-learning approaches streams have several limitations regards nature drifts, excessive required...

Customer journey analysis aims at understanding customer behavior both in the traditional offline setting and through online website visits. Particularly for latter, web analytics tools like Google Analytics maps have shown their usefulness, by being widely used companies. Nevertheless, they provide an oversimplified version of user addition to other limitations related narrow scope over cases. This paper contributes a novel approach overcome these applying process mining recommender systems...

Context prediction is an emerging topic in the fields of data mining and information management which both promising challenging. Predicting location mobile objects was a frequently tackled subtask context recent researches. For scenarios managing health persons, near future status persons at least equally important to predicting their location. We introduce this paper, best our knowledge, first method for next equipped with body sensors device. The suggested Prefix Span-based searches...

Accurate traffic flow prediction is an important tool to allow for more efficient use of networks. Current algorithms such as LSTM RNNs can be very successful in predicting regular flows, but often fail accurately predict the interesting irregularities flows. We propose OE-LSTM (Outlier-Enriched LSTM), a novel framework that focuses mainly on these irregular consider outliers within each stream and assume occur, certain set circumstances present cause deviations from pattern. After detecting...

Big Data Streams are very popular at now, as stirred-up by a plethora of modern applications such sensor networks, scientific computing tools, Web intelligence, social network analysis and mining so forth. Here, the main research issue consists in how to effectively efficiently extract useful knowledge from (streaming) big data, order support innovative data analytics platforms. To this end, clustering is well-known tool for extracting streams, also confirmed recent trends active literature....

Supporting sequential pattern mining from data streams is nowadays a relevant problem in the area of stream research. Actual proposals available literature are based on well-known PrefixSpan approach and are, indeed, able to effectively bound error discovered patterns. This foresees idea dividing target collection manageable chunks, i.e., pieces stream, order gain into effectiveness efficiency. Unfortunately, patterns chunks indeed introduce additional errors with respect basic application...

The analysis of the temporal evolution dynamic networks is a key challenge for understanding complex processes hidden in graph structured data. Graph rules capture such on level small subgraphs by describing frequently occurring structural changes within network. Existing rule discovery methods make restrictive assumptions change present networks. We propose EvoMine, frequent mining method that, first time, supports with edge insertions and deletions as well node relabelings. EvoMine defines...

Clustering of streaming sensor data aims at providing online summaries the observed stream. This task is mostly done under limited processing and storage resources. makes sensed stream speed (data per time) a sensitive restriction when designing clustering algorithms. Additionally, varying natural characteristic data, e.g. changing sampling rate upon detecting an event or for certain time. In such cases, most algorithms have to heavily restrict their model size that they can handle minimal...