FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2506116503> ?p ?o ?g. }

• W2506116503 abstract "In this thesis we study two important supervised learning settings: linear classifiers with a margin, and Multiple-Instance Learning, and provide novel results concerning the ability to learn in each of these settings. In supervised learning, the goal is to learn to classify objects into one of several classes, using only examples of objects, along with the class that they belong to (also termed their label). We focus on binary supervised learning, in which each object should be classified into one of two classes. As an example, consider the task of predicting whether a patient will present with diabetes, based on the patient’s blood test results. In this example, one class represents patients who will present with diabetes and the other class represents patients who will not present with diabetes. The learner is given a set of examples, where each example is constituted of the blood test results of a patient, along with information on whether this patient has presented with diabetes or not. We term this set of examples the training set, or the training sample. The training set is used by the learner to infer a classification rule, which can be used to predict whether a new patient will present with diabetes, based on this patient’s blood test results. The goal of the learner is to find a classification rule which is as accurate as possible in its predictions. An important measure of the effectiveness of learning is how many labeled examples are needed in order to achieve a certain degree of classification accuracy. The sample complexity of a learning problem is the size of a training set required to guarantee a given accuracy on this problem. Equivalently, it is the accuracy that can be guaranteed for the learner, given the size of the training sample. We distinguish between the sample complexity, which is a statistical measure of the difficulty of learning, and computational complexity, which measures the amount of computation required to implement a learning strategy. The “No free lunch” theorem for supervised learning [Wolpert and Macready, 1997] shows that no single supervised learning algorithm can provide a high-accuracy classification rule for all learning problems using the same sample size. In other words, the sample complexity of supervised learning without additional assumptions is unbounded. It follows that in order to have guarantees on learning, we need to consider more restricted classes of learning problems." @default.
• W2506116503 created "2016-08-23" @default.
• W2506116503 creator A5085310459 @default.
• W2506116503 date "2012-01-01" @default.
• W2506116503 modified "2023-09-26" @default.
• W2506116503 title "Partial Information and Distribution-Dependence in Supervised Learning Models" @default.
• W2506116503 cites W1482235099 @default.
• W2506116503 cites W1505356468 @default.
• W2506116503 cites W1516508599 @default.
• W2506116503 cites W1535599202 @default.
• W2506116503 cites W1539416424 @default.
• W2506116503 cites W1540386283 @default.
• W2506116503 cites W1542886316 @default.
• W2506116503 cites W1544144649 @default.
• W2506116503 cites W1547172162 @default.
• W2506116503 cites W1579249934 @default.
• W2506116503 cites W1586554030 @default.
• W2506116503 cites W1589919686 @default.
• W2506116503 cites W1601368642 @default.
• W2506116503 cites W1601541039 @default.
• W2506116503 cites W1676552347 @default.
• W2506116503 cites W1872991507 @default.
• W2506116503 cites W188867022 @default.
• W2506116503 cites W1971361630 @default.
• W2506116503 cites W1975846642 @default.
• W2506116503 cites W1981207991 @default.
• W2506116503 cites W1982618872 @default.
• W2506116503 cites W1988790447 @default.
• W2506116503 cites W1994592916 @default.
• W2506116503 cites W2003680476 @default.
• W2506116503 cites W2004915807 @default.
• W2506116503 cites W2010029425 @default.
• W2506116503 cites W2017753243 @default.
• W2506116503 cites W2019363670 @default.
• W2506116503 cites W2024052500 @default.
• W2506116503 cites W2029538739 @default.
• W2506116503 cites W2032210760 @default.
• W2506116503 cites W2040870580 @default.
• W2506116503 cites W2041615247 @default.
• W2506116503 cites W2042132284 @default.
• W2506116503 cites W2043856707 @default.
• W2506116503 cites W205342693 @default.
• W2506116503 cites W2061458158 @default.
• W2506116503 cites W2068484024 @default.
• W2506116503 cites W2074992691 @default.
• W2506116503 cites W2075567596 @default.
• W2506116503 cites W2081131431 @default.
• W2506116503 cites W2081177492 @default.
• W2506116503 cites W2084165935 @default.
• W2506116503 cites W2087347434 @default.
• W2506116503 cites W2092074494 @default.
• W2506116503 cites W2098300287 @default.
• W2506116503 cites W2098774896 @default.
• W2506116503 cites W2108745803 @default.
• W2506116503 cites W2110119381 @default.
• W2506116503 cites W2111377143 @default.
• W2506116503 cites W2112528242 @default.
• W2506116503 cites W2117049614 @default.
• W2506116503 cites W2119821739 @default.
• W2506116503 cites W2120419212 @default.
• W2506116503 cites W2121541286 @default.
• W2506116503 cites W2123469175 @default.
• W2506116503 cites W2129192653 @default.
• W2506116503 cites W2131043876 @default.
• W2506116503 cites W2144919274 @default.
• W2506116503 cites W2151554678 @default.
• W2506116503 cites W2154318594 @default.
• W2506116503 cites W2155904486 @default.
• W2506116503 cites W2156909104 @default.
• W2506116503 cites W2163474322 @default.
• W2506116503 cites W2165075008 @default.
• W2506116503 cites W2168494960 @default.
• W2506116503 cites W2222844749 @default.
• W2506116503 cites W2325684928 @default.
• W2506116503 cites W2402090481 @default.
• W2506116503 cites W2579923771 @default.
• W2506116503 cites W2964121570 @default.
• W2506116503 cites W2964270981 @default.
• W2506116503 cites W3013820469 @default.
• W2506116503 cites W3120740533 @default.
• W2506116503 cites W3121907121 @default.
• W2506116503 cites W3139996414 @default.
• W2506116503 cites W48137866 @default.
• W2506116503 cites W57549770 @default.
• W2506116503 cites W740415 @default.
• W2506116503 cites W1512533241 @default.
• W2506116503 hasPublicationYear "2012" @default.
• W2506116503 type Work @default.
• W2506116503 sameAs 2506116503 @default.
• W2506116503 citedByCount "4" @default.
• W2506116503 countsByYear W25061165032013 @default.
• W2506116503 countsByYear W25061165032014 @default.
• W2506116503 countsByYear W25061165032015 @default.
• W2506116503 crossrefType "journal-article" @default.
• W2506116503 hasAuthorship W2506116503A5085310459 @default.
• W2506116503 hasConcept C119857082 @default.
• W2506116503 hasConcept C12267149 @default.
• W2506116503 hasConcept C136389625 @default.
• W2506116503 hasConcept C151730666 @default.
• W2506116503 hasConcept C154945302 @default.

• next