Home About us Contact
|
Home About us Contact | ||
|
|
||
Set Only (set + only)
Selected AbstractsPrinciples of Proper Validation: use and abuse of re-sampling for validationJOURNAL OF CHEMOMETRICS, Issue 3-4 2010Kim H. Esbensen Abstract Validation in chemometrics is presented using the exemplar context of multivariate calibration/prediction. A phenomenological analysis of common validation practices in data analysis and chemometrics leads to formulation of a set of generic Principles of Proper Validation (PPV), which is based on a set of characterizing distinctions: (i) Validation cannot be understood by focusing on the methods of validation only; validation must be based on full knowledge of the underlying definitions, objectives, methods, effects and consequences,which are all outlined and discussed here. (ii) Analysis of proper validation objectives implies that there is one valid paradigm only: test set validation. (iii) Contrary to much contemporary chemometric practices (and validation myths), cross-validation is shown to be unjustified in the form of monolithic application of a one-for-all procedure (segmented cross-validation) on all data sets. Within its own design and scope, cross-validation is in reality a sub-optimal simulation of test set validation, crippled by a critical sampling variance omission, as it manifestly is based on one data set only (training data set). Other re-sampling validation methods are shown to suffer from the same deficiencies. The PPV are universal and can be applied to all situations in which the assessment of performance is desired: prediction-, classification-, time series forecasting-, modeling validation. The key element of PPV is the Theory of Sampling (TOS), which allow insight into all variance generating factors, especially the so-called incorrect sampling errors, which, if not properly eliminated, are responsible for a fatal inconstant sampling bias, for which no statistical correction is possible. In the light of TOS it is shown how a second data set (test set, validation set) is critically necessary for the inclusion of the sampling errors incurred in all ,future' situations in which the validated model must perform. Logically, therefore, all one data set re-sampling approaches for validation, especially cross-validation and leverage-corrected validation, should be terminated, or at the very least used only with full scientific understanding and disclosure of their detrimental variance omissions and consequences. Regarding PLS-regression, an emphatic call is made for stringent commitment to test set validation based on graphical inspection of pertinent t,u plots for optimal understanding of the X,Y interrelationships and for validation guidance. QSAR/QSAP forms a partial exemption from the present test set imperative with no generalization potential. Copyright © 2010 John Wiley & Sons, Ltd. [source] Omitting types in fuzzy logic with evaluated syntaxMLQ- MATHEMATICAL LOGIC QUARTERLY, Issue 3 2006Petra Murinová Abstract This paper is a contribution to the development of model theory of fuzzy logic in narrow sense. We consider a formal system Ev, of fuzzy logic that has evaluated syntax, i. e. axioms need not be fully convincing and so, they form a fuzzy set only. Consequently, formulas are provable in some general degree. A generalization of Gödel's completeness theorem does hold in Ev,. The truth values form an MV-algebra that is either finite or ,ukasiewicz algebra on [0, 1]. The classical omitting types theorem states that given a formal theory T and a set ,(x1, , , xn ) of formulas with the same free variables, we can construct a model of T which omits ,, i. e. there is always a formula from , not true in it. In this paper, we generalize this theorem for Ev,, that is, we prove that if T is a fuzzy theory and ,(x1, , , xn ) forms a fuzzy set , then a model omitting , also exists. We will prove this theorem for two essential cases of Ev,: either Ev, has logical (truth) constants for all truth values, or it has these constants for truth values from [0, 1] , , only. (© 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim) [source] Form, symmetry and packing of biomacromolecules.ACTA CRYSTALLOGRAPHICA SECTION A, Issue 3 2010The differentiation of the human rhinovirus into serotypes, all having very similar structures and the same architecture, is shown to be related to the packing of the viruses in the crystal and to its space-group symmetry. The molecular crystallographic properties (here described in terms of a molecular lattice ,M instead of the form lattice ,F considered in previous publications) appear to be compatible with the crystal structure and with the packing lattice ,P, introduced in Part I [Janner (2010). Acta Cryst. A66, 301,311]. On the basis of the enclosing forms of the capsid, a sphere packing is considered, where the spheres touch at kissing points. Residues of each of the four coat proteins (VP1, VP2, VP3, VP4), having a minimal distance from the kissing points, define a set of kissing point related (KPR) residues. In this set only four different residues occur, one for each coat protein, ordered into symmetric clusters {already classified in a previous publication [Janner (2006). Acta Cryst. A62, 270,286]} and indexed by neighbouring lattice points of ,P (or equivalently of ,M). The indexed KPR residues allow a fingerprint characterization of the five rhinovirus serotypes whose structures are known (HRV16, HRV14, HRV3, HRV2 and HRV1A). In the fingerprint they occur as internal (if inside the given capsid), as external (if belonging to the neighbouring viruses) or as a contact residue (if at a kissing point position). The same fingerprint, periodically extended, permits a coarse-grained reconstruction of the essential properties of the crystal packing, invariant with respect to the space group of the serotype. [source] Contribution of small insects to pollination of common buckwheat, a distylous cropANNALS OF APPLIED BIOLOGY, Issue 1 2009Hisatomo Taki Abstract Crop pollination by animals is an essential ecosystem service. Among animal-pollinated crops, distylous plants strongly depend on animal pollination. In distylous pollination systems, pollinator species are usually limited, although flowers of some distylous plants are visited by diverse animals. We studied the pollination biology of common buckwheat (Fagopyrum esculentum), a distylous crop mainly pollinated by honeybees and visited by many insect species, to evaluate the effects of non-honeybee species on pollination services. We focused on insects smaller than honeybees to determine their contribution to pollination. We applied pollination treatments with bags of coarse mesh to exclude flower visits by honeybees and larger insects and compared the seed set of bagged plants with that of untreated plants for pin and thrum flower morphs. We found a great reduction of seed set only in bagged pin flowers. We also confirmed that small insects, including ants, bees, wasps and flies, carried pin-morph pollen. These small insects transfer pollen from the short anthers of pin flowers to the short styles of thrum flowers, leading to sufficient seed set in thrum flowers. Consequently, small, non-honeybee insects have the potential to maintain at least half of the yield of this honeybee-dependent distylous crop. [source] Nucleosome Immobilization Strategies for Single-Pair FRET Microscopy,CHEMPHYSCHEM, Issue 14 2008Wiepke J. A. Koopmans Abstract All genomic transactions in eukaryotes take place in the context of the nucleosome, the basic unit of chromatin, which is responsible for DNA compaction. Overcoming the steric hindrance that nucleosomes present for DNA-processing enzymes requires significant conformational changes. The dynamics of these have been hard to resolve. Single-pair Fluorescence Resonance Energy Transfer (spFRET) microscopy is a powerful technique for observing conformational dynamics of the nucleosome. Nucleosome immobilization allows the extension of observation times to a limit set only by photobleaching, and thus opens the possibility of studying processes occurring on timescales ranging from milliseconds to minutes. It is crucial however, that immobilization itself does not introduce artifacts in the dynamics. Here we report on various nucleosome immobilization strategies, such as single-point attachment to polyethylene glycol (PEG) or surfaces coated with bovine serum albumin (BSA), and confinement in porous agarose or polyacrylamide gels. We compare the immobilization specificity and structural integrity of immobilized nucleosomes. A crosslinked star polyethylene glycol coating performs best with respect to tethering specificity and nucleosome integrity, and enables us to reproduce for the first time bulk nucleosome unwrapping kinetics in single nucleosomes without immobilization artifacts. [source] A new space and time sensor fusion method for mobile robot navigationJOURNAL OF FIELD ROBOTICS (FORMERLY JOURNAL OF ROBOTIC SYSTEMS), Issue 7 2004TaeSeok Jin To fully utilize the information from the sensors of mobile robot, this paper proposes a new sensor-fusion technique where the sample data set obtained at a previous instant is properly transformed and fused with the current data sets to produce a reliable estimate for navigation control. Exploration of an unknown environment is an important task for the new generation of mobile service robots. The mobile robots may navigate by means of a number of monitoring systems such as the sonar-sensing system or the visual-sensing system. Notice that in the conventional fusion schemes, the measurement is dependent on the current data sets only. Therefore, more sensors are required to measure a given physical parameter or to improve the reliability of the measurement. However, in this approach, instead of adding more sensors to the system, the temporal sequences of the data sets are stored and utilized for the purpose. The basic principle is illustrated by examples and the effectiveness is proved through simulations and experiments. The newly proposed STSF (space and time sensor fusion) scheme is applied to the navigation of a mobile robot in an environment using landmarks, and the experimental results demonstrate the effective performance of the system. © 2004 Wiley Periodicals, Inc. [source] Approximate decidability in euclidean spacesMLQ- MATHEMATICAL LOGIC QUARTERLY, Issue 1 2003Armin Hemmerling Abstract We study concepts of decidability (recursivity) for subsets of Euclidean spaces ,k within the framework of approximate computability (type two theory of effectivity). A new notion of approximate decidability is proposed and discussed in some detail. It is an effective variant of F. Hausdorff's concept of resolvable sets, and it modifies and generalizes notions of recursivity known from computable analysis, formerly used for open or closed sets only, to more general types of sets. Approximate decidability of sets can equivalently be expressed by computability of the characteristic functions by means of appropriately working oracle Turing machines. The notion fulfills some natural requirements and is hereditary under canonical embeddings of sets into spaces of higher dimensions. However, it is not closed under binary union or intersection of sets. We also show how the framework of resolvability and approximate decidability can be applied to investigate concepts of reducibility for subsets of Euclidean spaces. [source] | ||||||||||