Examinando por Autor "Bilbao Jayo, Aritz"

Mostrando 1 - 5 de 5
  • Miniatura
    Ítem
    Analysing centralities for organisational role inference in online social networks
    (Elsevier Ltd, 2021-03) Sánchez Corcuera, Rubén; Bilbao Jayo, Aritz; Zulaika Zurimendi, Unai; Almeida, Aitor
    The intensive use of Online Social Networks (OSN) nowadays has made users expose more information without realising it. Malicious users or marketing agencies are now able to infer information that is not published on OSNs by using data from targets friends to use for their benefit. In this paper, the authors present a generalisable method capable of deducing the roles of employees of an organisation using their Twitter relationships and the features of the graph from their organisation. The authors also conduct an extensive analysis of the node centralities to study their roles in the inference of the different classes proposed. Derived from the experiments and the ablation study conducted to the centralities, the authors conclude that the latent features of the graph along with the directed relationships perform better than previously proposed methods when classifying the role of the employees of an organisation. Additionally, to evaluate the method, the authors also contribute with a new dataset consisting of three directed graphs (one for each organisation) representing the relationships between the employees obtained from Twitter.
  • Miniatura
    Ítem
    A comparative analysis of human behavior prediction approaches in intelligent environments
    (MDPI, 2022-01-18) Almeida, Aitor; Bermejo Fernández, Unai ; Bilbao Jayo, Aritz ; Azkune Galparsoro, Gorka; Aguilera, Unai ; Emaldi, Mikel ; Dornaika, Fadi; Arganda-Carreras, Ignacio
    Behavior modeling has multiple applications in the intelligent environment domain. It has been used in different tasks, such as the stratification of different pathologies, prediction of the user actions and activities, or modeling the energy usage. Specifically, behavior prediction can be used to forecast the future evolution of the users and to identify those behaviors that deviate from the expected conduct. In this paper, we propose the use of embeddings to represent the user actions, and study and compare several behavior prediction approaches. We test multiple model (LSTM, CNNs, GCNs, and transformers) architectures to ascertain the best approach to using embeddings for behavior modeling and also evaluate multiple embedding retrofitting approaches. To do so, we use the Kasteren dataset for intelligent environments, which is one of the most widely used datasets in the areas of activity recognition and behavior modeling
  • Miniatura
    Ítem
    Embedding-based real-time change point detection with application to activity segmentation in smart home time series data
    (Elsevier Ltd, 2021-12-15) Bermejo Fernández, Unai; Almeida, Aitor; Bilbao Jayo, Aritz; Azkune Galparsoro, Gorka
    Human activity recognition systems are essential to enable many assistive applications. Those systems can be sensor-based or vision-based. When sensor-based systems are deployed in real environments, they must segment sensor data streams on the fly in order to extract features and recognize the ongoing activities. This segmentation can be done with different approaches. One effective approach is to employ change point detection (CPD) algorithms to detect activity transitions (i.e. determine when activities start and end). In this paper, we present a novel real-time CPD method to perform activity segmentation, where neural embeddings (vectors of continuous numbers) are used to represent sensor events. Through empirical evaluation with 3 publicly available benchmark datasets, we conclude that our method is useful for segmenting sensor data, offering significant better performance than state of the art algorithms in two of them. Besides, we propose the use of retrofitting, a graph-based technique, to adjust the embeddings and introduce expert knowledge in the activity segmentation task, showing empirically that it can improve the performance of our method using three graphs generated from two sources of information. Finally, we discuss the advantages of our approach regarding computational cost, manual effort reduction (no need of hand-crafted features) and cross-environment possibilities (transfer learning) in comparison to others
  • Miniatura
    Ítem
    From Political Manifestos to Social Networks
    (Universidad de Deusto, 2020-06-23) Bilbao Jayo, Aritz; Almeida, Aitor; Facultad de Ingeniería; Programa de Doctorado en Ingeniería para la Sociedad de la Información y Desarrollo Sostenible por la Universidad de Deusto
    Due to the rise of the social networks, political parties and politicians have found new ways of establishing their position on an issue apart from traditional political manifestos. From this phenomenon, a new research area has emerged, the automation of political discourse analysis on Social Networks. To do so, this PhD dissertation has taken advantage of a widely used content analysis methodology for political manifestos, The Manifesto Project. With annotated manifestos since 2001, this methodology uses a codification which allows the analysis of political parties policy preferences regarding 56 topics, providing the scientific community with parties’ policy positions derived from the content analysis. Therefore, this PhD dissertation focuses on two main tasks: firstly, to automate the annotation process of political manifestos, in order to facilitate that same process to political scientists and secondly, to use this model as a basis to perform a political discourse analysis on Twitter using the previously mentioned Manifesto Project's methodology. To do so, we have taken advantage of two types of contextual information available in the two circumstances of the application of this research work: manifestos and Twitter. The first contextual data is what has been said previously, in the case of election manifestos the previous phrase or statement, and on twitter the preceding tweet. The second contextual information is which political party is the sender of the statement. Regarding the use of contextual information in order to improve manifestos automated classification, we have improved state of the art results in 4 out of 7 languages. With regard to Tweets' classification, we can affirm that annotated manifestos can be used as complementary data for this task, being the fine-tuned model with annotated tweets the best performing one. Moreover, contextual information does also improve the performance of the models when tweets are classified. Using this approach, we have analysed the 2016 United States presidential elections on Twitter.
  • Miniatura
    Ítem
    A multicenter randomized trial for quality of life evaluation by non-invasive intelligent tools during post-curative treatment follow-up for head and neck cancer: clinical study protocol
    (Frontiers Media S.A., 2023-01-31) Cavalieri, Stefano; Vener, Claudia; LeBlanc, Marissa; Lopez-Perez, Laura; Fico, Giuseppe; Resteghini, Carlo; Monzani, Dario; Marton, Giulia; Pravettoni, Gabriella; Moreira-Soares, Mauricio; Filippidou, Despina Elizabeth; Almeida, Aitor; Bilbao Jayo, Aritz ; Mehanna, Hisham; Singer, Susanne; Thomas, Steve; Lacerenza, Luca; Manfuso, Alfonso; Copelli, Chiara; Mercalli, Franco; Frigessi, Arnoldo; Martinelli, Elena; Licitra, Lisa; Estevez-Priego, Estefania
    Patients surviving head and neck cancer (HNC) suffer from high physical, psychological, and socioeconomic burdens. Achieving cancer-free survival with an optimal quality of life (QoL) is the primary goal for HNC patient management. So, maintaining lifelong surveillance is critical. An ambitious goal would be to carry this out through the advanced analysis of environmental, emotional, and behavioral data unobtrusively collected from mobile devices. The aim of this clinical trial is to reduce, with non-invasive tools (i.e., patients’ mobile devices), the proportion of HNC survivors (i.e., having completed their curative treatment from 3 months to 10 years) experiencing a clinically relevant reduction in QoL during follow-up. The Big Data for Quality of Life (BD4QoL) study is an international, multicenter, randomized (2:1), open-label trial. The primary endpoint is a clinically relevant global health-related EORTC QLQ-C30 QoL deterioration (decrease ≥10 points) at any point during 24 months post-treatment follow-up. The target sample size is 420 patients. Patients will be randomized to be followed up using the BD4QoL platform or per standard clinical practice. The BD4QoL platform includes a set of services to allow patients monitoring and empowerment through two main tools: a mobile application installed on participants’ smartphones, that includes a chatbot for e-coaching, and the Point of Care dashboard, to let the investigators manage patients data. In both arms, participants will be asked to complete QoL questionnaires at study entry and once every 6 months, and will undergo post-treatment follow up as per clinical practice. Patients randomized to the intervention arm (n=280) will receive access to the BD4QoL platform, those in the control arm (n=140) will not. Eligibility criteria include completing curative treatments for non-metastatic HNC and the use of an Android-based smartphone. Patients undergoing active treatments or with synchronous cancers are excluded. Clinical Trial Registration: ClinicalTrials.gov, identifier (NCT05315570).