The Experts below are selected from a list of 11697 Experts worldwide ranked by ideXlab platform
Zahidul Islam - One of the best experts on this subject based on the ideXlab platform.
-
a differentially private Random Decision Forest using reliable signal to noise ratios
Australasian Joint Conference on Artificial Intelligence, 2015Co-Authors: Sam Fletcher, Zahidul IslamAbstract:When dealing with personal data, it is important for data miners to have algorithms available for discovering trends and patterns in the data without exposing people’s private information. Differential privacy offers an enforceable definition of privacy that can provide each individual in a dataset a guarantee that their personal information is no more at risk than it would be if their data was not in the dataset at all. By using mechanisms that achieve differential privacy, we propose a Decision Forest algorithm that uses the theory of Signal-to-Noise Ratios to automatically tune the algorithm’s parameters, and to make sure that any differentially private noise added to the results does not outweigh the true results. Our experiments demonstrate that our differentially private algorithm can achieve high prediction accuracy.
Celikic Minja - One of the best experts on this subject based on the ideXlab platform.
-
Fourier Transform Infrared (FT-IR) and Laser Ablation Inductively Coupled Plasma–Mass Spectrometry (LA-ICP-MS) Imaging of Cerebral Ischemia : Combined Analysis of Rat Brain Thin Cuts Toward Improved Tissue Classification
2018Co-Authors: Balbekova Anna, Lohninger Hans, Van Tilborg, Geralda A.f., Dijkhuizen, Rick M., Bonta Maximilian, Limbeck Andreas, Lendl Bernhard, Al-saad, Khalid A., Ali Mohamed, Celikic MinjaAbstract:Microspectroscopic techniques are widely used to complement histological studies. Due to recent developments in the field of chemical imaging, combined chemical analysis has become attractive. This technique facilitates a deepened analysis compared to single techniques or side-by-side analysis. In this study, rat brains harvested one week after induction of photothrombotic stroke were investigated. Adjacent thin cuts from rats’ brains were imaged using Fourier transform infrared (FT-IR) microspectroscopy and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The LA-ICP-MS data were normalized using an internal standard (a thin gold layer). The acquired hyperspectral data cubes were fused and subjected to multivariate analysis. Brain regions affected by stroke as well as unaffected gray and white matter were identified and classified using a model based on either partial least squares discriminant analysis (PLS-DA) or Random Decision Forest (RDF) algorithms. The RDF algorithm demonstrated the best results for classification. Improved classification was observed in the case of fused data in comparison to individual data sets (either FT-IR or LA-ICP-MS). Variable importance analysis demonstrated that both molecular and elemental content contribute to the improved RDF classification. Univariate spectral analysis identified biochemical properties of the assigned tissue types. Classification of multisensor hyperspectral data sets using an RDF algorithm allows access to a novel and in-depth understanding of biochemical processes and solid chemical allocation of different brain regions
Celikic M. - One of the best experts on this subject based on the ideXlab platform.
-
Fourier Transform Infrared (FT-IR) and Laser Ablation Inductively Coupled Plasma–Mass Spectrometry (LA-ICP-MS) Imaging of Cerebral Ischemia: Combined Analysis of Rat Brain Thin Cuts Toward Improved Tissue Classification
'SAGE Publications', 2018Co-Authors: Balbekova A., Lohninger H., Van Tilborg G.a.f., Dijkhuizen R.m., Bonta M., Limbeck A., Lendl B., Al-saad K.a., Ali M., Celikic M.Abstract:Microspectroscopic techniques are widely used to complement histological studies. Due to recent developments in the field of chemical imaging, combined chemical analysis has become attractive. This technique facilitates a deepened analysis compared to single techniques or side-by-side analysis. In this study, rat brains harvested one week after induction of photothrombotic stroke were investigated. Adjacent thin cuts from rats’ brains were imaged using Fourier transform infrared (FT-IR) microspectroscopy and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The LA-ICP-MS data were normalized using an internal standard (a thin gold layer). The acquired hyperspectral data cubes were fused and subjected to multivariate analysis. Brain regions affected by stroke as well as unaffected gray and white matter were identified and classified using a model based on either partial least squares discriminant analysis (PLS-DA) or Random Decision Forest (RDF) algorithms. The RDF algorithm demonstrated the best results for classification. Improved classification was observed in the case of fused data in comparison to individual data sets (either FT-IR or LA-ICP-MS). Variable importance analysis demonstrated that both molecular and elemental content contribute to the improved RDF classification. Univariate spectral analysis identified biochemical properties of the assigned tissue types. Classification of multisensor hyperspectral data sets using an RDF algorithm allows access to a novel and in-depth understanding of biochemical processes and solid chemical allocation of different brain regions.This work was supported by MEIBio doctoral project of TU Wien and by the funding from the Austrian FFG within project 84247. This work was co-funded by NPRP grant no. NPRP-5-381-3-101 from the Qatar National Research Fund (a member of The Qatar Foundation).Scopu
Sam Fletcher - One of the best experts on this subject based on the ideXlab platform.
-
a differentially private Random Decision Forest using reliable signal to noise ratios
Australasian Joint Conference on Artificial Intelligence, 2015Co-Authors: Sam Fletcher, Zahidul IslamAbstract:When dealing with personal data, it is important for data miners to have algorithms available for discovering trends and patterns in the data without exposing people’s private information. Differential privacy offers an enforceable definition of privacy that can provide each individual in a dataset a guarantee that their personal information is no more at risk than it would be if their data was not in the dataset at all. By using mechanisms that achieve differential privacy, we propose a Decision Forest algorithm that uses the theory of Signal-to-Noise Ratios to automatically tune the algorithm’s parameters, and to make sure that any differentially private noise added to the results does not outweigh the true results. Our experiments demonstrate that our differentially private algorithm can achieve high prediction accuracy.
Balbekova Anna - One of the best experts on this subject based on the ideXlab platform.
-
Fourier Transform Infrared (FT-IR) and Laser Ablation Inductively Coupled Plasma–Mass Spectrometry (LA-ICP-MS) Imaging of Cerebral Ischemia : Combined Analysis of Rat Brain Thin Cuts Toward Improved Tissue Classification
2018Co-Authors: Balbekova Anna, Lohninger Hans, Van Tilborg, Geralda A.f., Dijkhuizen, Rick M., Bonta Maximilian, Limbeck Andreas, Lendl Bernhard, Al-saad, Khalid A., Ali Mohamed, Celikic MinjaAbstract:Microspectroscopic techniques are widely used to complement histological studies. Due to recent developments in the field of chemical imaging, combined chemical analysis has become attractive. This technique facilitates a deepened analysis compared to single techniques or side-by-side analysis. In this study, rat brains harvested one week after induction of photothrombotic stroke were investigated. Adjacent thin cuts from rats’ brains were imaged using Fourier transform infrared (FT-IR) microspectroscopy and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The LA-ICP-MS data were normalized using an internal standard (a thin gold layer). The acquired hyperspectral data cubes were fused and subjected to multivariate analysis. Brain regions affected by stroke as well as unaffected gray and white matter were identified and classified using a model based on either partial least squares discriminant analysis (PLS-DA) or Random Decision Forest (RDF) algorithms. The RDF algorithm demonstrated the best results for classification. Improved classification was observed in the case of fused data in comparison to individual data sets (either FT-IR or LA-ICP-MS). Variable importance analysis demonstrated that both molecular and elemental content contribute to the improved RDF classification. Univariate spectral analysis identified biochemical properties of the assigned tissue types. Classification of multisensor hyperspectral data sets using an RDF algorithm allows access to a novel and in-depth understanding of biochemical processes and solid chemical allocation of different brain regions
