Articles publicats (IIIA)

Generation of Individualized Synthetic Data for Augmentation of the Type 1 Diabetes Data Sets Using Deep Learning Models

Thu, 30 Jun 2022 00:00:00 GMT

Generation of Individualized Synthetic Data for Augmentation of the Type 1 Diabetes Data Sets Using Deep Learning Models Noguer, Josep; Contreras, Ivan; Mujahid, Omer; Beneyto Tantiña, Aleix; Vehí, Josep In this paper, we present a methodology based on generative adversarial network architecture to generate synthetic data sets with the intention of augmenting continuous glucose monitor data from individual patients. We use these synthetic data with the aim of improving the overall performance of prediction models based on machine learning techniques. Experiments were performed on two cohorts of patients suffering from type 1 diabetes mellitus with significant differences in their clinical outcomes. In the first contribution, we have demonstrated that the chosen methodology is able to replicate the intrinsic characteristics of individual patients following the statistical distributions of the original data. Next, a second contribution demonstrates the potential of synthetic data to improve the performance of machine learning approaches by testing and comparing different prediction models for the problem of predicting nocturnal hypoglycemic events in type 1 diabetic patients. The results obtained for both generative and predictive models are quite encouraging and set a precedent in the use of generative techniques to train new machine learning models

Impact of Use Frequency of a Mobile Diabetes Management App on Blood Glucose Control: Evaluation Study

Fri, 01 Mar 2019 00:00:00 GMT

Impact of Use Frequency of a Mobile Diabetes Management App on Blood Glucose Control: Evaluation Study Vehí, Josep; Regincós i Isern, Jordi; Parcerisas Albés, Adrià; Calm i Puig, Remei; Contreras, Ivan Technology has long been used to carry out self-management as well as to improve adherence to treatment in people with diabetes. However, most technology-based apps do not meet the basic requirements for engaging patients. Objective: This study aimed to evaluate the effect of use frequency of a diabetes management app on glycemic control. Methods: Overall, 2 analyses were performed. The first consisted of an examination of the reduction of blood glucose (BG) mean, using a randomly selected group of 211 users of the SocialDiabetes app (SDA). BG levels at baseline, month 3, and month 6 were calculated using the intercept of a regression model based on data from months 1, 4, and 7, respectively. In the second analysis, the impact of low and high BG risk was examined. A total of 2692 users logging SDA ≥5 days/month for ≥6 months were analyzed. The highest quartile regarding low blood glucose index (LBGI) and high blood glucose index (HBGI) at baseline (t1) was selected (n=74 for group A; n=440 for group B). Changes in HBGI and LBGI at month 6 (t2) were analyzed. Results: For analysis 1, baseline BG results for type 1 diabetes mellitus (T1DM) groups A and B were 213.61 (SD 31.57) mg/dL and 206.43 (SD 18.65) mg/dL, respectively, which decreased at month 6 to 175.15 (SD 37.88) mg/dL and 180.6 (SD 40.47) mg/dL, respectively. For type 2 diabetes mellitus (T2DM), baseline BG was 218.77 (SD 40.18) mg/dL and 232.55 (SD 46.78) mg/dL, respectively, which decreased at month 6 to 160.51 (SD 39.32) mg/dL and 173.14 (SD 52.81) mg/dL for groups A and B, respectively. This represents a reduction of estimated A1c (eA1c) of approximately 1.3% (P<.001) and 0.9% (P=.001) for T1DM groups A and B, respectively, and 2% (P<.001) for both A and B T2DM groups, respectively. For analysis 2, T1DM baseline LBGI values for groups A and B were 5.2 (SD 3.9) and 4.4 (SD 2.3), respectively, which decreased at t2 to 3.4 (SD 3.3) and 3.4 (SD 1.9), respectively; this was a reduction of 34.6% (P=.005) and 22.7% (P=.02), respectively. Baseline HBGI values for groups A and B were 12.6 (SD 4.3) and 10.6 (SD 4.03), respectively, which decreased at t2 to 9.0 (SD 6.5) and 8.6 (SD 4.7), respectively; this was a reduction of 30% (P=.001) and 22% (P=.003), respectively. Conclusions: A significant reduction in BG was found in all groups, independent of the use frequency of the app. Better outcomes were found for T2DM patients. A significant reduction in LBGI and HBGI was found in all groups, regardless of the use frequency of the app. LBGI and HBGI indices of both groups tend to have similar values after 6 months of app use

Experimental study of semiactive VSC techniques for vehicle vibration reduction

Tue, 01 Jan 2013 00:00:00 GMT

Experimental study of semiactive VSC techniques for vehicle vibration reduction Pozo, Francesc; Zapateiro de la Hoz, Mauricio Fabián; Acho, Leonardo; Vidal, Yolanda; Luo, Ningsu Semiactive vehicle suspension with magnetorheological dampers is a promising technology for improving the passenger comfort of a ground vehicle. In this work we consider the problem of vibration suppression in suspension systems of an experimental platform called semiactive suspension system (SAS). The system under study is equipped with a magnetorheological damper which is used as a semiactive nonlinear device. Three different variable structure controllers are proposed and experimentally tested. Experimental results show that the vibration of the suspension system is well controlled with the semiactive VSC. Furthermore, it is found that the implementation of the controllers implies a better performance than the use of a purely passive device. Despite the performance of all the controllers is comparable, the one based on the updating term yields the best results among them all

Power re-allocation for reducing contracted electric power costs

Sun, 15 Feb 2015 00:00:00 GMT

Power re-allocation for reducing contracted electric power costs Torrent-Fontbona, Ferran; López Ibáñez, Beatriz Electric bills consist of a cost related to the consumed energy and a cost related to the maximum demanded power. This latter part usually accounts for approximately 25-40% of the bill. Demanded power by big consumers is measured in real time and electric companies highly penalise them if the maximum demanded power (along the billing period) exceeds the contracted power by the consumer. In this paper we propose a new method that, given the demanded power of close consumers for a time window (power profile), power costs are reduced by re-allocating the demanded power among consumers in order to keep all of them below or equal to their contracted power. We also propose and analyse some strategies to set a preference when not all power profiles can be kept below the contracted power. We tested this method using real-based simulated power profiles of eight different business buildings located in Girona and the power cost reduction achieved reached approximately 20%