22 Hits in 3.4 sec

Analysing acoustic model changes for active learning in automatic speech recognition

Chenhao Wu, Raymond W. M. Ng, Oscar Saz Torralba, Thomas Hain
2017 2017 International Conference on Systems, Signals and Image Processing (IWSSIP)  
A preliminary attempt in active learning proves the effectiveness of GSD selection criterion over random selection, shedding light on its prospective use.  ...  It is found that GSD provide hints for predicting data transcription quality.  ...  Apart from speaker f 6, using the GSD selection method with 10% relabelled data gave almost as good performance as using the random selection method with double (20%) relabelled data. VIII.  ... 
doi:10.1109/iwssip.2017.7965609 dblp:conf/iwssip/WuNSH17 fatcat:sybspv53y5amxge7e7rvk7ll74

Using Stock Prices as Ground Truth in Sentiment Analysis to Generate Profitable Trading Signals [article]

Ellie Birbeck, Dave Cliff
2018 arXiv   pre-print
A Bayesian classifier is used to generate stock predictions, which are input to an automated trading algorithm.  ...  Previous work in this field investigates how the true sentiment of text (i.e. positive or negative opinions) can be used for financial predictions, based on the assumption that sentiments expressed online  ...  to estimate the linear dependency between two variables (here this refers to the class and the term of a document), and as with χ 2 , this method of feature selection returns univariate scores for the  ... 
arXiv:1811.02886v1 fatcat:kbysx5t3x5cnpbma6lgv6roije

Using Stock Prices as Ground Truth in Sentiment Analysis to Generate Profitable Trading Signals

Ellie Birbeck, Dave Cliff
2018 2018 IEEE Symposium Series on Computational Intelligence (SSCI)  
to estimate the linear dependency between two variables (here this refers to the class and the term of a document), and as with χ 2 , this method of feature selection returns univariate scores for the  ...  Fig. 6 : 6 Profits trading FB stock with Methods A, B, C. TABLE I : I Validation set accuracy of each model with varying feature selection methods and subset sizes.  ... 
doi:10.1109/ssci.2018.8628841 dblp:conf/ssci/BirbeckC18 fatcat:bzckzjzas5fvdkqqelnjenmmma

Evaluation of a novel therapeutic education programme for people with alcohol use disorder in France: a mixed-methods intervention study protocol (ETHER)

Saskia Antwerpes, Marie Costa, Marion Coste, Morgane Bureau, Gwenaelle Maradan, Christophe Cutarella, Jacques Leloutre, Olivier Riccobono-Soulier, Sophie Hedoire, Elodie Frot, Fabienne Vernier, Stéphanie Vassas-Goyard (+3 others)
2022 Harm Reduction Journal  
Linear regression models will be used to assess the impact of the TPE programme on changes in the primary and secondary outcomes, while adjusting for other correlates and confounders.  ...  set of harm reduction criteria for use in future research.  ...  (D) Repeat analysis B and C for secondary outcomes using linear regression models.  ... 
doi:10.1186/s12954-021-00587-0 pmid:35012570 pmcid:PMC8751002 fatcat:i2wloh6wtfgmzmjcrlyjsxndpu

Mining linguistic tone patterns with symbolic representation

2016 Proceedings of the 14th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology  
We discuss the potential of SPM using time-series mining techniques in future works.  ...  to possibilities of using a wide range of algorithms that require discrete valued data.  ...  The seed pattern can be selected using example top K-motifs extracted from the corpus, or using a user-supplied pattern (numeric, symbolic data points, or audio files).  ... 
doi:10.18653/v1/w16-2001 dblp:conf/sigmorphon/Zhang16 fatcat:ynrf47sb2zaxzdjocgcni4ypma

The Relationships between Vocabulary Learning Strategies and Vocabulary Size among Adult Spanish Foreign Language Learners

Dieter A. Waldvogel
2013 Journal of Language Teaching and Research  
The data were analyzed through quantitative methods using two measuring instruments: a vocabulary learning questionnaire used to discover students' VLS preferences, and a vocabulary test used to estimate  ...  The aim of this study is to investigate the relationships between the type of vocabulary learning strategies (VLS) adult Spanish foreign language (FL) students use at different levels of proficiency and  ...  We would also like to thank the instructors who generously helped us collect the data during their classes and all the students who were willing to participate in the study.  ... 
doi:10.4304/jltr.4.2.209-219 fatcat:tlhqs6wo6jb4jjak2oi7oubnpy

A study on deep neural network acoustic model adaptation for robust far-field speech recognition

Seyedmahdad Mirsamadi, John H. L. Hansen
2015 Interspeech 2015   unpublished
Using various adaptation strategies, relative word error rate improvements of up to 16% could be obtained on the single-channel task of the recent Aspire challenge.  ...  Linear transformations closer to the input layer provided consistently better results compared to the other methods.  ...  Although these methods have originally been developed for speaker adaptation, they are also able to provide significant improvements when the mismatch is due to environmental distortions.  ... 
doi:10.21437/interspeech.2015-525 fatcat:dxhubexl55csho5srjqknancqu

Working Memory Training for Healthy Older Adults: The Role of Individual Characteristics in Explaining Short- and Long-Term Gains

Erika Borella, Elena Carbone, Massimiliano Pastore, Rossana De Beni, Barbara Carretti
2017 Frontiers in Human Neuroscience  
Results: Linear mixed models confirmed the overall efficacy of the training, in the short-term at least, and some maintenance effects.  ...  in the training (criterion task)-predict the short-and long-term specific gains and transfer effects of a verbal WM training for older adults.  ...  RESULTS Linear Mixed Models Linear mixed effects (LME) models were used to analyze the data because, unlike the more classical and frequentlyused methods, they enable estimates to be adjusted for repeat  ... 
doi:10.3389/fnhum.2017.00099 pmid:28381995 pmcid:PMC5360719 fatcat:3q5422rz4rdbfa5i5mf4crnq3e

Controlling Recurrent Neural Networks by Conceptors [article]

Herbert Jaeger
2017 arXiv   pre-print
predict the cur- rent input.  ...  Methods section).  ... 
arXiv:1403.3369v3 fatcat:4gaiy3cikvervdjvsmp47b2byy

IRRODL Volume 15, Number 6

Varios Authors
2014 International Review of Research in Open and Distance Learning  
However, as a generally affordable method of post-secondary education delivery, MOOCs place the developing countries at the centre of universal access to HE.  ...  Their study used Rogers' diffusion of innovation model and compared the perceptions of mobile LMS users and nonusers.  ...  Researchers used a mixed-method analysis to evaluate responses to the selected questions.  ... 
doi:10.19173/irrodl.v15i6.2063 fatcat:tspb3bvb45cajgko7vhrh736ca

Abstracts of the 19th European Conference on Eye Movements 2017 [article]

Radach, Ralph; University Of Wuppertal, Vorstius, Christian; University Of Wuppertal, Hofmann (Eds.), Markus J.; University Of Wuppertal, Deubel, Heiner; University Of Munich
Using a General Linear Model, these potentials can be identified separately. Methods.  ...  Methods. We manipulate noise and adaptation using predictability of the retinal projection of the stimulus.  ...  Using the same target words in a lexical decision task (Experiment 3b), we did obtain a significant effect for root frequency, but it was qualified by a significant interaction with letter s tring lexicality  ... 
doi:10.16910/jemr.10.6 fatcat:3vdhqhh3ibbjvge5pnqgljq24a

Analysis of Executables: Benefits and Challenges Edited by Seminar topics

Andy King, Alan Mycroft, Thomas Reps, Axel Simon, A Gb, Ac King@kent, Uk, Axel Simon, Andy King
Adpating these methods to binaries require static identification of procedure 'call' and 'ret' instructions.  ...  BEST can be used during testing to predict program failures, or during debugging to replay program failures.  ...  A Formal ARM Model and Its Use This tool has been used in an extension of the L4.verified project which proved functional correctness of the seL4 microkernel.  ... 

An Analysis of the Argumentative Writing Skills of Academically Underprepared College Students

Beth Jillian Chase
Descriptive statistics and ordinal logistic regression were used to analyze a total of 112 writing samples.  ...  Using an archival data set, a sample of argumentative essays written by community college developmental (remedial) education students was analyzed.  ...  Conventional foods that are produced is usually has the use of methods to fertilize. See convential foods use fertilizers to promote plant growth.  ... 
doi:10.7916/d8hm5gdz fatcat:5xnxjr3hwzh5fad6klcoqg22ra

Symmetric Cryptography (Dagstuhl Seminar 12031) Learning in Multiobjective Optimization (Dagstuhl Seminar 12041) Analysis of Executables: Benefits and Challenges (Dagstuhl Seminar 12051)

Robert Hirschfeld, Shriram Krishnamurthi, Jan, Verónica Becher, Laurent Bienvenu, Rodney Downey, Elvira Mayordomo, Frederik Armknecht, Stefan Lucks, Bart Preneel, Phillip Rogaway, Salvatore Greco (+6 others)
2012 Foundations for Scripting Languages   unpublished
Acknowledgments Many thanks to the Dagstuhl office and its helpful and patient staff; huge thanks to the organizers of the previous seminars in the series for setting us up for success; and thanks to all  ...  Moreover, a linear value function seems to predict choices for bi-criteria problems quite well. This ability to predict is independent of whether the value function is diagnosed linear or not.  ...  BEST can be used during testing to predict program failures, or during debugging to replay program failures.  ... 

Joint Discourse-aware Concept Disambiguation and Clustering

Angela Petra Fahrni
Markov logic combines first order logic with probabilities and allows us to concisely formalize these interdependencies.  ...  As an inventory, we use Wikipedia. We compare our approach to other approaches and show that it achieves state-of-the-art results.  ...  Different contexts can be differently weighted and by using feature selection methods different contexts can be selected.  ... 
doi:10.11588/heidok.00020737 fatcat:vhljgiqbrbcwtpjxce6k4vcp2a
« Previous Showing results 1 — 15 out of 22 results