FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W32396075> ?p ?o ?g. }

• W32396075 abstract "The gene expression data of a sample is a vector containing the expression levels of many genes in the sample measured simultaneously by DNA microarray. From the viewpoint of pattern recognition, the task of cancer class ification based on gene expression data is a pattern classif ication problem, and the feature vector for the classification is the gene expression vector. However, this problem is an extremely difficult one for many methods, since the feature dimension is usually very high (usually several thousands), and the training samples (samples whose correct classification is known and thus plays the role as supervisors) are usually very scarce (say, around 100 known samples or less). If directly working in this high dimensional space with limited samples, most conventional pattern recognition algorithms may not work well. Some algorithm (algorithms that involve matrix inversion operation) may not be able to arrive at a solution when the number of samples is less than the dimensionality. For others that can achieve a solution, it may not be able to work well on samples other than that used for training. This is called the generalization problem in pattern recognition and machine learning. Thus, most previous methods for cancer classification based on gene expression data starts with a feature selection procedure. For example, Golub et al defined a metric for evaluating the correlation of a gene with a classification scheme, thus determining whether the gene is relevant or not (Golub et al, 1999; Slonim et al, 2000). Obviously, this kind of strategy does not take possible correlation and co-action among the genes into consideration. Unless it can be proven that the genes are statistically independent with each other (or orthogonal to each other), the result is far from optimal. We believe that due to the complex (and unknown) relationship among the genes, it should be better to start with the analysis of the full data set. For this purpose, we should be able to design a good classifier with all the candidate genes. Then the most relevant genes for this classification can be discovered by evaluating the subset of genes that contributes mostly in this classifier. We choose SVM or Support Vector Machine as our classifier, due to its supposed good performance on extremely scarce samples in high-dimensional space." @default.
• W32396075 created "2016-06-24" @default.
• W32396075 creator A5015401462 @default.
• W32396075 creator A5081242795 @default.
• W32396075 date "2001-01-01" @default.
• W32396075 modified "2023-09-23" @default.
• W32396075 title "Recursive Sample Classification and Gene Selection based on SVM: Method and Software Description #" @default.
• W32396075 cites W1486089539 @default.
• W32396075 cites W1714373868 @default.
• W32396075 cites W2096059361 @default.
• W32396075 cites W2108728387 @default.
• W32396075 cites W2109363337 @default.
• W32396075 cites W2119821739 @default.
• W32396075 cites W2122599407 @default.
• W32396075 cites W2148603752 @default.
• W32396075 cites W2149298154 @default.
• W32396075 cites W2156909104 @default.
• W32396075 cites W99903558 @default.
• W32396075 hasPublicationYear "2001" @default.
• W32396075 type Work @default.
• W32396075 sameAs 32396075 @default.
• W32396075 citedByCount "3" @default.
• W32396075 countsByYear W323960752013 @default.
• W32396075 crossrefType "journal-article" @default.
• W32396075 hasAuthorship W32396075A5015401462 @default.
• W32396075 hasAuthorship W32396075A5081242795 @default.
• W32396075 hasConcept C104317684 @default.
• W32396075 hasConcept C111030470 @default.
• W32396075 hasConcept C119857082 @default.
• W32396075 hasConcept C12267149 @default.
• W32396075 hasConcept C124101348 @default.
• W32396075 hasConcept C134306372 @default.
• W32396075 hasConcept C148483581 @default.
• W32396075 hasConcept C150194340 @default.
• W32396075 hasConcept C153180895 @default.
• W32396075 hasConcept C154945302 @default.
• W32396075 hasConcept C177148314 @default.
• W32396075 hasConcept C2984324147 @default.
• W32396075 hasConcept C33923547 @default.
• W32396075 hasConcept C41008148 @default.
• W32396075 hasConcept C54355233 @default.
• W32396075 hasConcept C70518039 @default.
• W32396075 hasConcept C83665646 @default.
• W32396075 hasConcept C8415881 @default.
• W32396075 hasConcept C86803240 @default.
• W32396075 hasConceptScore W32396075C104317684 @default.
• W32396075 hasConceptScore W32396075C111030470 @default.
• W32396075 hasConceptScore W32396075C119857082 @default.
• W32396075 hasConceptScore W32396075C12267149 @default.
• W32396075 hasConceptScore W32396075C124101348 @default.
• W32396075 hasConceptScore W32396075C134306372 @default.
• W32396075 hasConceptScore W32396075C148483581 @default.
• W32396075 hasConceptScore W32396075C150194340 @default.
• W32396075 hasConceptScore W32396075C153180895 @default.
• W32396075 hasConceptScore W32396075C154945302 @default.
• W32396075 hasConceptScore W32396075C177148314 @default.
• W32396075 hasConceptScore W32396075C2984324147 @default.
• W32396075 hasConceptScore W32396075C33923547 @default.
• W32396075 hasConceptScore W32396075C41008148 @default.
• W32396075 hasConceptScore W32396075C54355233 @default.
• W32396075 hasConceptScore W32396075C70518039 @default.
• W32396075 hasConceptScore W32396075C83665646 @default.
• W32396075 hasConceptScore W32396075C8415881 @default.
• W32396075 hasConceptScore W32396075C86803240 @default.
• W32396075 hasOpenAccess W32396075 @default.
• W32396075 hasRelatedWork W1494922272 @default.
• W32396075 hasRelatedWork W1587362683 @default.
• W32396075 hasRelatedWork W1630964756 @default.
• W32396075 hasRelatedWork W2087684630 @default.
• W32396075 hasRelatedWork W2097839764 @default.
• W32396075 hasRelatedWork W2108728387 @default.
• W32396075 hasRelatedWork W2109363337 @default.
• W32396075 hasRelatedWork W2137476312 @default.
• W32396075 hasRelatedWork W2143426320 @default.
• W32396075 hasRelatedWork W2147246240 @default.
• W32396075 hasRelatedWork W2158825237 @default.
• W32396075 hasRelatedWork W99903558 @default.
• W32396075 hasRelatedWork W1543148020 @default.
• W32396075 hasRelatedWork W1558237107 @default.
• W32396075 hasRelatedWork W1578454553 @default.
• W32396075 hasRelatedWork W1596386844 @default.
• W32396075 hasRelatedWork W183258039 @default.
• W32396075 hasRelatedWork W1834552780 @default.
• W32396075 hasRelatedWork W1912225830 @default.
• W32396075 hasRelatedWork W1948606734 @default.
• W32396075 isParatext "false" @default.
• W32396075 isRetracted "false" @default.
• W32396075 magId "32396075" @default.
• W32396075 workType "article" @default.