1,721,079 research outputs found
From Event Logs to Subprocesses: Supporting the Analysis of Unstructured Processes
No full textOggigiorno molte organizzazioni gestiscono i propri processi attraverso sistemi informatici che registrano i dati delle esecuzioni dei processi sui cosiddetti event log. Le tecniche di Process Mining (PM) usano tali log per rappresentare, monitorare e migliorare i corrispondenti processi. Una problematica ancora aperta riguarda l’applicazione delle tecniche di PM a log di processi non strutturati, che tende a restituire modelli complessi, chiamati modelli spaghetti, che forniscono un limitato supporto all’analisi. In questa tesi si intende affrontare tale problematica proponendo un approccio per estrarre da event log di processi non strutturati i sottoprocessi più rilevanti, al posto di modelli completi. Dapprima si costruisce un Instance Graph (IG) per ogni traccia del log, rappresentante il flusso di esecuzione della relativa esecuzione di processo. Poi si applica una tecnica di Frequent Subgraph Mining per estrarre i sottografi (cioè i sottoprocessi) più rilevanti dall’insieme di IG. Poiché gli IG costruiti per processi non strutturati sono spesso modelli imprecisi, si introduce una procedura di riparazione per migliorarne la qualità. I risultati sperimentali ottenuti su event log sintetici e reali attestano che tale procedura migliora significativamente la qualità degli IG ottenuti e, quindi, dei sottoprocessi estratti. Gli esperimenti condotti su event log reali dimostrano anche la capacità dell’approccio di evidenziare aspetti significativi dei rispettivi processi. Nell’ultima parte della tesi si esplorano gli aspetti legati alla collaborazione, attraverso applicazioni e casi di studio. Si analizzano sistemi di supporto all’innovazione e l’applicazione dell’approccio all’analisi di schemi di innovazione di imprese a rete. In seguito, si affronta l’estrazione delle pratiche di collaborazione di un team, esplorando anche i problemi dovuti alla mancanza di un sistema informatico dedicato. Vantaggi e limiti vengono valutati su casi di studio sintetici e reali
Emerging challenges in legal informatics from machine learning to LLMs - Preface to the proceedings of the 1st PLC workshop
No full textThe integration of Artificial Intelligence techniques, machine learning and large language models into legal informatics offers innovative potential, from enhancing legal research efficiency to supporting legal reasoning. These advancements introduce significant challenges, including issues related to data privacy, bias in legal datasets, and the interpretability of complex algorithms in legal contexts. Emerging challenges involve reliability, fairness, and ethical considerations in AI-driven legal applications. The research contributions presented at a recent workshop on Processes, Law and Compliance aim to deepen these issues for the development of AI applications in the field of legal informatics.</p
ESub: Exploration of Subgraphs A tool for exploring models generated by Graph Mining algorithms
No full textIn this demo we introduce ESub, a tool aimed at visualizing the outcome provided by a frequent subgraph mining algorithm, i.e. SUBDUE. Such a tool has been developed as a supporting tool for a methodology we proposed in previous works for analyzing unstructured processes, based on the use of graphs. By exploiting graphs-based techniques, it is possible to provide the user with a different perspective on a process, where only the most relevant subprocesses (i.e., subgraphs) are displayed, rather than the complete, end-to-end process schema, which often results very chaotic in unstructured domains. Our tool allows the user to visualize and interact with such subgraphs. Furthermore, it allows for visualizing the original graphs of the set, and compress them by means of the most relevant subgraphs, in order to obtain a simplified view of the overall process
Understanding Knowlegde-Intensive Processes: from Traces to Instance Graphs
No full textEnterprise information systems, while support daily activities, typically collect data on executed processes in event logs. These data describe the temporal sequence in which activities were carried out, hiding possible parallelism and other control flows. Representing the structure of each process execution in the form of an Instance Graph, enables managers to discover valuable knowledge on enterprise behaviors. In this work, we describe BIG4ProM, a tool which implements the Building Instance Graph (BIG) algorithm. BIG4ProM exploits filtering Process Discovery algorithms implemented in ProM in order to return the set of instance graphs related to the given event log. The plug-in is conceived to support both expert and standard users
Approximation of the gradient of the error probability for vector quantizers
No full textVector Quantizers (VQ) can be exploited for classification. In particular the gradient of the error probability performed by a VQ with respect to the position of its code vectors can be formally derived, hence the optimum VQ can be theoretically found. Unfortunately, this equation is of limited use in practice, since it relies on the knowledge of the class conditional probability distributions. In order to apply the method to real problems where distributions are unknown, a stochastic approximation has been previously proposed to derive a practical learning algorithm. In this paper we relax some of the assumptions underlying the original proposal and study the advantages of the resulting algorithm by both synthetic and real case studies
Pattern Discovery from Innovation Processes
Full text linkInnovation management and promotion has become one of the most important topics in the Literature about business and executive decision support. In particular, the relationship between innovation and collaboration, both intra- and inter-organization, is gaining an increasing attention in many works, for example in the Open Innovation research field. Innovation activities, especially those that involve collaboration, are typically not structured; they don’t follow a predefined scheme or procedure and are influenced by multiple factors, for
instance the individual behaviour, that makes it difficult to apply classical methods of process analysis. In this paper we describe a methodology to discover significant and recurrent patterns in innovation activities, that can be used to support and improve such kind of processes. To evaluate our approach we conducted a set of experiments on a synthetic dataset, which contains a set of traces of innovation activities generated from some abstract templates, drew with the aim to model the typical ways in which innovation is carried on
A semi-automatic methodology for the design of performance monitoring systems
No full textIn the present work, we propose a methodology for the design of a strategic support information system, aimed both at monitoring enterprise daily activities and at supporting decision making by means of Key Performance Indicators (KPIs). In particular, given a set of requested KPIs and the schemas of available data sources, our approach aims at identifying the subset of requested KPIs that can be actually computed over the sources. The KPIs are represented by means of an ontology, over which proper reasoning functionalities have been imple- mented. Both such automatic functionalities and interactions with ex- perts are required in order to map ontology concepts to schema elements
Towards a Customizable User-Centered Model for Data Analytics
No full textEvidence-based governance and e-democracy both rely on the capability to analyze aggregated and statistical data. Recent studies report that existing analysis tools were never fully embraced by managers mainly because of their complexity for many analytical use cases. This is even more true for citizens, that do not have full control over underlying data and analysis models. In the present work, we propose an innovative user-centered approach for data analytics, that facilitates the interaction of users with statistical and aggregated measures, i.e. indicators. We provide an overview of the framework, discussing its main components and functionalities. In particular we focus on an ontology representing both atomic and compound indicators, that
are provided with a calculation formula. We show how such a logic-based representation of indicators allows the implementation of powerful, automatic reasoning services, capable to provide a valuable support to users for performing analysis tasks
Semantic Process Mining for Ambient Assisted Living
No full textAmbient Assisted Living (AAL) systems are aimed to assist elderly people and enhance their autonomy, by monitoring their health, supporting their daily activities and so on. AAL tools are employed for several purposes, e.g. medication management, social isolation prevention, fall detection. In this work, we focus on the analysis of daily activities of monitored people and, in particular, on the detection of common patterns of daily activities. These patterns allow to understand the habitual behavior of monitored people, that is a valuable knowledge both in order to enhance the support provided to elders in performing their activities and to be able to quickly detect unexpected or dangerous situations. However, AAL tools usually return data at a very low level of detail, analyzing which too detailed patterns are inferred, which are of scarce support for the human analyst. To address this issue, in this work, we discuss the application of a combined methodology based on the combination of semantic techniques and multidimensional analysis paradigm to allow the analyst to switch to the desired level of granularity and to consider different process perspective, thus enhancing the analysis
- …
