Derivative-augmented features as a dynamic model for time-series. Baggenstoss, P. M. In *2015 23rd European Signal Processing Conference (EUSIPCO)*, pages 958-962, Aug, 2015.

Paper doi abstract bibtex

Paper doi abstract bibtex

In the field of automatic speech recognition (ASR), it is common practice to augment features with time-derivatives, which we call derivative-augmented features (DAF). Although the method is effective for modeling the dynamic behavior of features and produces signiicantly lower clas-siication error, it violates the assumption of conditional independence of the observations. The traditional approach is to ignore the problem (simply apply the mathematical approach that assumes independence). In this paper, we take an alternative approach in which we still use the same mathematical approach as before, but calculate a correction factor by integrating out the redundant dimensions. This makes it possible to compare and combine a DAF PDF and a non-DAF PDF. We conduct experiments to demonstrate the usefulness of the approach.

@InProceedings{7362525, author = {P. M. Baggenstoss}, booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)}, title = {Derivative-augmented features as a dynamic model for time-series}, year = {2015}, pages = {958-962}, abstract = {In the field of automatic speech recognition (ASR), it is common practice to augment features with time-derivatives, which we call derivative-augmented features (DAF). Although the method is effective for modeling the dynamic behavior of features and produces signiicantly lower clas-siication error, it violates the assumption of conditional independence of the observations. The traditional approach is to ignore the problem (simply apply the mathematical approach that assumes independence). In this paper, we take an alternative approach in which we still use the same mathematical approach as before, but calculate a correction factor by integrating out the redundant dimensions. This makes it possible to compare and combine a DAF PDF and a non-DAF PDF. We conduct experiments to demonstrate the usefulness of the approach.}, keywords = {correlation methods;hidden Markov models;signal classification;speech recognition;time series;derivative-augmented features;time-series;automatic speech recognition;ASR;time-derivatives;DAF;classification error;mathematical approach;correction factor calculation;redundant dimensions;hidden Markov model;HMM;Hidden Markov models;Markov processes;Europe;Signal processing;Indexes;Probability density function;Feature extraction;PDF estimation;feature derivatives;HMM}, doi = {10.1109/EUSIPCO.2015.7362525}, issn = {2076-1465}, month = {Aug}, url = {https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570103093.pdf}, }

Downloads: 0

{"_id":"qukyvcduFfEfj5oR6","bibbaseid":"baggenstoss-derivativeaugmentedfeaturesasadynamicmodelfortimeseries-2015","downloads":0,"creationDate":"2016-01-08T22:33:14.240Z","title":"Derivative-augmented features as a dynamic model for time-series","author_short":["Baggenstoss, P. M."],"year":2015,"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/Roznn/EUSIPCO/main/eusipco2015url.bib","bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["P.","M."],"propositions":[],"lastnames":["Baggenstoss"],"suffixes":[]}],"booktitle":"2015 23rd European Signal Processing Conference (EUSIPCO)","title":"Derivative-augmented features as a dynamic model for time-series","year":"2015","pages":"958-962","abstract":"In the field of automatic speech recognition (ASR), it is common practice to augment features with time-derivatives, which we call derivative-augmented features (DAF). Although the method is effective for modeling the dynamic behavior of features and produces signiicantly lower clas-siication error, it violates the assumption of conditional independence of the observations. The traditional approach is to ignore the problem (simply apply the mathematical approach that assumes independence). In this paper, we take an alternative approach in which we still use the same mathematical approach as before, but calculate a correction factor by integrating out the redundant dimensions. This makes it possible to compare and combine a DAF PDF and a non-DAF PDF. We conduct experiments to demonstrate the usefulness of the approach.","keywords":"correlation methods;hidden Markov models;signal classification;speech recognition;time series;derivative-augmented features;time-series;automatic speech recognition;ASR;time-derivatives;DAF;classification error;mathematical approach;correction factor calculation;redundant dimensions;hidden Markov model;HMM;Hidden Markov models;Markov processes;Europe;Signal processing;Indexes;Probability density function;Feature extraction;PDF estimation;feature derivatives;HMM","doi":"10.1109/EUSIPCO.2015.7362525","issn":"2076-1465","month":"Aug","url":"https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570103093.pdf","bibtex":"@InProceedings{7362525,\n author = {P. M. Baggenstoss},\n booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)},\n title = {Derivative-augmented features as a dynamic model for time-series},\n year = {2015},\n pages = {958-962},\n abstract = {In the field of automatic speech recognition (ASR), it is common practice to augment features with time-derivatives, which we call derivative-augmented features (DAF). Although the method is effective for modeling the dynamic behavior of features and produces signiicantly lower clas-siication error, it violates the assumption of conditional independence of the observations. The traditional approach is to ignore the problem (simply apply the mathematical approach that assumes independence). In this paper, we take an alternative approach in which we still use the same mathematical approach as before, but calculate a correction factor by integrating out the redundant dimensions. This makes it possible to compare and combine a DAF PDF and a non-DAF PDF. We conduct experiments to demonstrate the usefulness of the approach.},\n keywords = {correlation methods;hidden Markov models;signal classification;speech recognition;time series;derivative-augmented features;time-series;automatic speech recognition;ASR;time-derivatives;DAF;classification error;mathematical approach;correction factor calculation;redundant dimensions;hidden Markov model;HMM;Hidden Markov models;Markov processes;Europe;Signal processing;Indexes;Probability density function;Feature extraction;PDF estimation;feature derivatives;HMM},\n doi = {10.1109/EUSIPCO.2015.7362525},\n issn = {2076-1465},\n month = {Aug},\n url = {https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570103093.pdf},\n}\n\n","author_short":["Baggenstoss, P. M."],"key":"7362525","id":"7362525","bibbaseid":"baggenstoss-derivativeaugmentedfeaturesasadynamicmodelfortimeseries-2015","role":"author","urls":{"Paper":"https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570103093.pdf"},"keyword":["correlation methods;hidden Markov models;signal classification;speech recognition;time series;derivative-augmented features;time-series;automatic speech recognition;ASR;time-derivatives;DAF;classification error;mathematical approach;correction factor calculation;redundant dimensions;hidden Markov model;HMM;Hidden Markov models;Markov processes;Europe;Signal processing;Indexes;Probability density function;Feature extraction;PDF estimation;feature derivatives;HMM"],"metadata":{"authorlinks":{}},"downloads":0},"search_terms":["derivative","augmented","features","dynamic","model","time","series","baggenstoss"],"keywords":["correlation methods;hidden markov models;signal classification;speech recognition;time series;derivative-augmented features;time-series;automatic speech recognition;asr;time-derivatives;daf;classification error;mathematical approach;correction factor calculation;redundant dimensions;hidden markov model;hmm;hidden markov models;markov processes;europe;signal processing;indexes;probability density function;feature extraction;pdf estimation;feature derivatives;hmm"],"authorIDs":[],"dataSources":["eov4vbT6mnAiTpKji","knrZsDjSNHWtA9WNT"]}