Search

Search Funnelback University

Search powered by Funnelback
41 - 90 of 112 search results for Economics test |u:mobile-systems.cl.cam.ac.uk where 27 match all words and 85 match some words.
  1. Results that match 1 of 2 words

  2. Investigating Domain-agnostic Performance in Activity Recognition…

    https://mobile-systems.cl.cam.ac.uk/papers/hasca22.pdf
    5 Sep 2022: Human activity recognition (HAR) models suffer significant performance degradation when faced with data heterogeneity(device, users, environments) at test time. ... This work presents the case for training models which are domain-agnostic, i.e., that
  3. Stress Inference from Abdominal Sounds using Machine Learning Erika…

    https://mobile-systems.cl.cam.ac.uk/papers/embc22-bondareva.pdf
    22 Apr 2022: Remain seated. Complete the daily task. randomDo anything. Meditation(relaxing task). Stroop test(stressful task). ... Moleman, H. G. van Steenis, et al., “Characterizationof stress reactions to the stroop color word test,” Pharmacol.
  4. IMChew: Chewing Analysis using Earphone Inertial Measurement Units

    https://mobile-systems.cl.cam.ac.uk/papers/bodysys24-yang.pdf
    2 May 2024: 5.1 Chewing DetectionTable 2 presents the overall performance of different classi-fiers for chewing detection when evaluated with an 80/20train-test split. ... Nonetheless, oursolution performs well under both evaluation methods. (a) 80/20 Train-test
  5. Conditional Neural ODE Processes for Individual Disease Progression…

    https://mobile-systems.cl.cam.ac.uk/papers/kdd23.pdf
    1 Jun 2023: The. smaller the -Γ𝑝𝑏(), the better the predicted disease progressionmatching the test labels. ... Only the initial audio sample is used asthe context vector during the test phase.
  6. UR2M: Uncertainty and Resource-Aware Event Detection on…

    https://mobile-systems.cl.cam.ac.uk/papers/percom24.pdf
    5 Feb 2024: After preprocessing, we obtained 92,502 total. event training samples (90%) and 10,278 test samples (10%). ... quantification method generating multiple test samples by. applying data augmentation techniques through a single model.
  7. Towards Adversarial Robustness with Early Exit Ensembles Lorena…

    https://mobile-systems.cl.cam.ac.uk/papers/embc22-qendro.pdf
    22 Apr 2022: All datasets are split into 80%/10%/10%train/validation/test maintaining class proportions. Eachdataset is paired with a different architecture: FCNet [17](fully-convolutional 5-layer network) for ECG and ... Additionally, we provide anal-ysis on
  8. imwut20a-sub7831-cam-i26

    https://mobile-systems.cl.cam.ac.uk/papers/contauth.pdf
    9 Nov 2020: The datasets were divided into ve parts:training, validation, test 1, test 2, and test 3. ... We used the remaining 10 subjects to cast attacks during the testing session: test 1.
  9. KDD__19_Spathis_et_al_nocopyright

    https://mobile-systems.cl.cam.ac.uk/papers/KDD19Spathis.pdf
    20 May 2019: Learned patterns of individual neurons. We now inspecthow the individual neurons of the decoder layer re as we passthe test-set through them. ... Compari-son with the actual mood variability (b). all user-weeks in the test set are very dierent.
  10. CTG: A Connectivity Trace Generator for Testing thePerformance of ...

    https://mobile-systems.cl.cam.ac.uk/papers/esec07.pdf
    5 Aug 2008: Real traces. Trace Analyser. Trace generator. CTG. Range Variation. Connectivity traces (Test cases). ... 22] Z. Wang, S. Elbaum, and D. Rosenblum. Automated generation ofcontext-aware tests.
  11. EmotionSense: A Mobile Phones based Adaptive Platform for…

    https://mobile-systems.cl.cam.ac.uk/papers/Ubicomp10.pdf
    29 Jun 2010: We will present the results of some of these tests asmethodological example in the next section. ... A separate,held-out dataset was used to test the accuracy of the speakerrecognition component.
  12. Exploring Automatic Diagnosis of COVID-19 from Crowdsourced…

    https://mobile-systems.cl.cam.ac.uk/papers/KDD_covid_19.pdf
    4 Aug 2020: Figure 3 (b) shows the most frequent symptoms of usersdeclaring a positive COVID test. ... Note that we used augmented samples only for training (the test set waskept intact).
  13. IEEE JOURNAL OF BIOMEDICAL AND HEALTH IN-FORMATICS 1…

    https://mobile-systems.cl.cam.ac.uk/papers/JBHI24Xia.pdf
    8 Apr 2024: The datasets used encompassvarious modalities, and all of them exhibit severe classimbalance, making them ideal test beds for the evaluation. ... FOLDS WITH DIFFERENT SEEDS. #TEST IS THE TESTING SIZE. C IS THE NUMBER OF CLASSES AND D IS THE INPUT DATA
  14. Yawning Detection using Earphone Inertial Measurement Units

    https://mobile-systems.cl.cam.ac.uk/papers/smartwear23.pdf
    4 Oct 2023: learning and ensure fair test-ing. ... A set ofusers making up around 25% of all data was selected, then80% of the data from these users was selected (20% of thetotal data) to be the test set.
  15. OESense: Employing Occlusion Effect for In-ear Human Sensing

    https://mobile-systems.cl.cam.ac.uk/papers/mobisys21.pdf
    25 May 2021: Otherwise, the user needs to adjustthe earbud and perform the fit test again. ... For each iter-ation in the leave-one-out tests, we include different amounts ofdata from the testing subject for training and test on the rest.
  16. Writing on the Clean Slate:Implementing a Socially-Aware Protocol in…

    https://mobile-systems.cl.cam.ac.uk/papers/aoc08.pdf
    17 Apr 2008: 4.4 Functional Testing. In order to perform a functional test of the integration ofGently in Haggle, we have set up a testbed of four desk-top computers equipped with 108 ... We randifferent test to observe if the asynchronous delivery pro-cess was
  17. 1 On the Effectiveness of an OpportunisticTraffic Management System…

    https://mobile-systems.cl.cam.ac.uk/papers/its11.pdf
    13 Sep 2011: Tests were run ona 4 km 7 km urban area in Tokyo, area which contained85 intersections. ... We performed two tests:One-hop gossiping: In this experiment, each vehicle gos-.
  18. Experience in deploying wearable devices for office analytics

    https://mobile-systems.cl.cam.ac.uk/papers/cscw16.pdf
    29 Mar 2016: At the end of the deployment, we asked our participants tocomplete a Big-5 personality test in order to capture theirpersonality traits.
  19. Proceedings on Privacy Enhancing Technologies ..; .. (..):1–17…

    https://mobile-systems.cl.cam.ac.uk/papers/pets2018-manousakas.pdf
    15 Mar 2018: method tests graphicalmodels of varying orders and selects the optimal orderby balancing the model complexity and the explanatorypower of observations. ... To do so, the adver-sary computes the pairwise distances between trainingmobility networks and
  20. Exploiting Place Features in Link Prediction onLocation-based Social…

    https://mobile-systems.cl.cam.ac.uk/papers/kdd2011.pdf
    30 May 2011: we test what predic-tion performance can be achieved by using only one featureclass with respect to the full model. ... used. Results averagedthe three snapshots and over 20 different randomtraining and test sets.
  21. Enabling On-Device Smartphone GPU basedTraining: Lessons Learned…

    https://mobile-systems.cl.cam.ac.uk/papers/perfail22.pdf
    4 Feb 2022: These details are described next. Model Architecture. Having failed at beating the CPU inthe experiments devised above, we move on to trying outdifferent architectures to test how size affects the
  22. Exploiting Foursquare and Cellular Data to InferUser Activity in ...

    https://mobile-systems.cl.cam.ac.uk/papers/mdm2013.pdf
    20 Mar 2013: We test the prediction accuracyof supervised learning algorithms by exposing them to amulti-class classification scenario, where a single urbanactivity has to be elected amongst a set of those. ... They arefollowed by Shopping and Parks and Outdoor areas,
  23. CROSS-DEVICE FEDERATED LEARNING FOR MOBILE HEALTH DIAGNOSTICS:A FIRST …

    https://mobile-systems.cl.cam.ac.uk/papers/ICASSP23-Xia.pdf
    24 Mar 2023: After data cleaning (i.e., excluding non-English speak-ers, samples without COVID-19 test results and poor audio qualitysamples), there are 482 users with positive status and 2, 478 userswith
  24. Passive mobile sensing and psychological traits for large scale mood …

    https://mobile-systems.cl.cam.ac.uk/papers/pervasivehealth19.pdf
    20 May 2019: test [25], and then we transform theseselected features with Principal Component Analysis (PCA) [30]to obtain feature combinations with the maximum variance. ... Therefore, we create training and test sets fromdisjoint user splits, making sure that weeks
  25. Evaluating Context Information Predictabilityfor Autonomic…

    https://mobile-systems.cl.cam.ac.uk/papers/acc2006.pdf
    29 Feb 2008: In order to test this, we consider the normalised valueof z(t) by dividing it by 100.
  26. Predictive Resource Scheduling in Computational Grids∗ Clovis…

    https://mobile-systems.cl.cam.ac.uk/papers/ipdps07.pdf
    29 Feb 2008: Dino Sally. Workload generator. Jobs Jobs. (With prediction). Figure 5. Test-bed Architecture. ... Figure 5 shows an overview of the experimental test-bed.The test-bed comprises two Condor clusters of 23 nodeseach.
  27. Mining User Mobility Features for Next PlacePrediction in…

    https://mobile-systems.cl.cam.ac.uk/papers/icdm2012.pdf
    20 Oct 2012: Results: We are now presenting the prediction resultsobtained when we train and test the two supervised learningmodels. ... In this case the authors propose asupervised learning model based on the places visited by auser’s friends and they test via
  28. 1 Talking Places: Modelling and Analysing LinguisticContent in…

    https://mobile-systems.cl.cam.ac.uk/papers/socialcom12_bauer.pdf
    21 Jul 2012: We average the results of 20 runs (or“particles”) on the same test set to obtain the final result. ... InTable VIII, we report the total log-likelihood divided by thenumber of tokens in the test set.
  29. IMPROVING FEATURE GENERALIZABILITY WITH MULTITASK LEARNING IN…

    https://mobile-systems.cl.cam.ac.uk/papers/icassp22-ma.pdf
    22 Apr 2022: Similarly,GSC dataset is split to 5-3-3-3-3-3. For both datasets, the (train,test, validation) splitting ratio is set to (0.7, 0.2, 0.1).
  30. Evolution of a Location-based Online Social Network:Analysis and…

    https://mobile-systems.cl.cam.ac.uk/papers/imc2012.pdf
    7 Sep 2012: Given real dataabout the evolution of a network, one can test the extent. ... 7.2 EvaluationTo test our model we take the real network at the begin-.
  31. What Will You Do for the Rest of the Day? An Approach to Continuous…

    https://mobile-systems.cl.cam.ac.uk/papers/imwut1819.pdf
    27 Feb 2019: What Will You Do for the Rest of the Day? An Approach to. Continuous Trajectory Prediction. AMIN SADRI, FLORA D. SALIM, YONGLI REN, and WEI SHAO, RMIT University, AustraliaJOHN C. KRUMM, Microsoft Research, USACECILIA MASCOLO, University of Cambridge
  32. LifeLearner: Hardware-Aware Meta Continual Learning System for…

    https://mobile-systems.cl.cam.ac.uk/papers/sensys23.pdf
    23 Oct 2023: 10 𝑆𝑟𝑒𝑒𝑎𝑟𝑠𝑎𝑙 = {𝑆𝑟𝑒𝑒𝑎𝑟𝑠𝑎𝑙, 𝐵𝑖𝑡𝑃𝑄𝑐𝑜𝑚𝑝𝑟𝑒𝑠𝑠 (𝑆𝑙𝑎𝑡𝑒𝑛𝑡 )}11 𝑆𝑡𝑒𝑠𝑡 = T 𝑆𝑡𝑟𝑎𝑖𝑛 // Held-out test set12
  33. Measuring Urban Social Diversity Using Interconnected Geo-Social…

    https://mobile-systems.cl.cam.ac.uk/papers/www2016.pdf
    1 Feb 2016: ABSTRACTLarge metropolitan cities bring together diverse individuals, creat-ing opportunities for cultural and intellectual exchanges, which canultimately lead to social and economic enrichment. ... Predicting gentrifi-cation of neighbourhoods could help
  34. Adaptive Routing for Intermittently Connected Mobile Ad Hoc Networks…

    https://mobile-systems.cl.cam.ac.uk/papers/wowmom05.pdf
    29 Feb 2008: discrete event simulator [14]. In order to obtain credible re-sults and to test the peculiar characteristics of our protocol,it was also necessary for us to develop a new group ... We implemented a simplified version of the DSDV pro-tocol [10] in order
  35. wifiparking.dvi

    https://mobile-systems.cl.cam.ac.uk/papers/mobicom2013.pdf
    23 Jul 2013: Figure 3: Energy consumption of network location. consumption of both of our test devices while performinga passive scan for Wi-Fi access points. ... we include these traces in our evaluation to test for falsepositives for driving detection.
  36. METIS: Exploring mobile phone sensing offloading for efficiently…

    https://mobile-systems.cl.cam.ac.uk/papers/percom2013.pdf
    24 Jan 2013: IV. BENCHMARKS. In this section we present the evaluation of the gainthreshold offloading scheme through micro-benchmark tests.We compared the energy performance of the scheme with thetwo schemes described in ... In these cases the local phone sensor
  37. An Ad Hoc Mobility Model Foundedon Social Network Theory ...

    https://mobile-systems.cl.cam.ac.uk/papers/mswim04.pdf
    29 Feb 2008: Using psycho-logical tests it is probable that the importance of a rela-tionship, such as a friendship, will be valued differently bythe different individuals involved; in our modelisation, thiswould lead
  38. A Study of Bluetooth Low Energy Performance forHuman Proximity ...

    https://mobile-systems.cl.cam.ac.uk/papers/percom17.pdf
    11 Jan 2017: Additionally, theavailability of our prototype allowed us to test the impact ofthese parameters on a large scale deployment. ... Our approach instead was to deploy our prototype, giving usgreater flexibility, and then test different combinations of
  39. 2 PERVASIVE computing Published by the IEEE CS n ...

    https://mobile-systems.cl.cam.ac.uk/papers/ieeepervasive2013.pdf
    12 Jun 2013: emotionsense evaluationWe evaluated the EmotionSense sys-tem through several offline microben-chmark tests and a deployment with 18 participants, who recorded their emo-tions in a daily diary. ... Our sensing framework will au-tomate these controls while
  40. A Random Walk Around the City:New Venue Recommendation in ...

    https://mobile-systems.cl.cam.ac.uk/papers/socialcom12_noulas.pdf
    21 Jul 2012: Thus, we couple a training dataset to a test data set which belongs to the following and non-overlapping temporal period. ... We note that users with no check-ins in the test set are notincluded in the performance evaluation.
  41. Epcast: Controlled Dissemination in Human-based Wireless Networks…

    https://mobile-systems.cl.cam.ac.uk/papers/epcast.pdf
    24 Mar 2008: beguaranteed. 5.2 Experimental Evaluation. Description of the Simulation In order to test the performance of thesetechniques, we defined a square simulation area with a side of 1 km and a
  42. Earbuds_Percom21 (1)

    https://mobile-systems.cl.cam.ac.uk/papers/percom21.pdf
    30 Apr 2021: in-the-wild use test.
  43. 978-1-4244-2575-4/08/$20.00 c©2008 IEEE Seal-2-Seal: A Delay-Tolerant …

    https://mobile-systems.cl.cam.ac.uk/papers/mass08.pdf
    17 Jul 2008: Small-scale tests have been performed on theTmote Sky sensor platform [9] to verify that the protocolworks on real hardware, and larger scale deployments areplanned for the future.
  44. SCAR: Context-aware Adaptive Routing inDelay Tolerant Mobile Sensor…

    https://mobile-systems.cl.cam.ac.uk/papers/dtmn06.pdf
    29 Feb 2008: We are also in the process ofporting the algorithm on top of Telos Motes running Con-tiki [7] in order to evaluate SCAR on a real test bed.
  45. Mobile Health with Head-Worn Devices: Challenges and Opportunities

    https://mobile-systems.cl.cam.ac.uk/papers/earable-survey22.pdf
    5 Sep 2022: Theseexisting forms have already been accepted by the users andhave been proved to withstand the test of time, therefore thetransition to a smart device is easier.
  46. ICGes v1

    https://mobile-systems.cl.cam.ac.uk/papers/icges.pdf
    29 Feb 2008: In order to test and validate the usefulness of. AEGIS as well as its ease of use (for facilitators.
  47. SociableSense: Exploring the Trade-offs of AdaptiveSampling and…

    https://mobile-systems.cl.cam.ac.uk/papers/mobicom2011.pdf
    9 Jul 2011: We perform several micro-benchmark tests tofine-tune and evaluate the performance of these mechanismsand we show that the adaptive sampling and computationdistribution schemes balance trade-offs among accuracy, en-ergy, ... NEP isa stand-alone test and
  48. Multimodal Deep Learning for Activity and Context Recognition

    https://mobile-systems.cl.cam.ac.uk/papers/ubicomp2018-radu.pdf
    21 Dec 2017: applied to wearable modeling tasks. Within this variety of deep learning, we are also the first to test thisarchitecture in-conjunction with convolutional layers (relative to feed-forward fully-connected DNN ... of nodes) for each task, guided by the
  49. Far from the Eyes, Close on the Web: Impact ...

    https://mobile-systems.cl.cam.ac.uk/papers/wosn12-kaltenbrunner.pdf
    18 Jun 2012: Economic Geography,46:234–240, 1970. [21] S. Traverso, K. Huguenin, V. Trestian, I.
  50. Discovering Latent Patterns of Urban Cultural Interactions in WeChat…

    https://mobile-systems.cl.cam.ac.uk/papers/kdd18.pdf
    18 Jun 2018: 2017. Cultural investment and urban socio-economic development: ageosocial network approach. Royal Society open science 4, 9 (2017), 170413.
  51. Evolution and Sustainability of aWildlife Monitoring Sensor Network…

    https://mobile-systems.cl.cam.ac.uk/papers/sensys10.pdf
    1 Sep 2010: The network status messages, which contain the radio on-time at each node, allowed us to derive the average radio dutycycle of each node over the test period. ... Figure 7. Results from the network test: (a) Box plot showing the distribution of radio

Search history

Recently clicked results

Recently clicked results

Your click history is empty.

Recent searches

Recent searches

Your search history is empty.