Скачать презентацию

Идет загрузка презентации. Пожалуйста, подождите

Презентация была опубликована 2 года назад пользователемВалентин Зайцев

1 Elastic Maps for Data Analysis Alexander Gorban, Leicester with Andrei Zinovyev, Paris

2 Plan of the talk INTRODUCTION Two paradigms for data analysis: statistics and modelling Clustering and K-means Self Organizing Maps PCA and local PCA

3 Plan of the talk 1. Principal manifolds and elastic maps The notion of of principal manifold (PM) Constructing PMs: elastic maps Adaptation and grammars 2. Application technique Projection and regression Maps and visualization of functions 3. Implementation and examples

4 Two basic paradigms for data analysis Data set Statistical Analysis Data Modelling

5 Statistical Analysis Existence of a Probability Distribution; Statistical Hypothesis about Data Generation; Verification/Falsification of Hypothesises about Hidden Properties of Data Distribution

6 Data Modelling We should find the Best Model for Data description; We know the Universe of Models; We know the Fitting Criteria; Learning Errors and Generalization Errors analysis for the Model Verification Universe of models

7 Example: Simplest Clustering

8 K-means algorithm Centers y (i) Data points x (j) 1)Minimize U for given {K (i) } (find centers); 2)Minimize U for given {y (i) } (find classes); 3)If {K (i) } change, then go to step 1.

9 Centers can be lines, manifolds,… with the same algorithm 1 st Principal components + mean points for classes instead of simplest means

10
SOM - Self Organizing Maps Set of nodes is a finite metric space with distance d(N,M); 0) Map set of nodes into dataspace N f 0 (N); 1) Select a datapoint X (random); 2) Find a nearest f i (N) (N=N X ); 3) f i+1 (N) = f i (N) +w i (d(N, N X ))(X- f i (N)), where w i (d) (0

11 PCA and Local PCA The covariance matrix is positive definite (X q are datapoints) Principal components: eigenvectors of the covariance matrix: The local covariance matrix (w is a positive cutting function) The field of principal components: eigenvectors of the local covariance matrix, e i (y). Trajectories of these vector-fields present geometry of local data structure.

12 A top secret: the difference between two basic paradigms is not crucial (Almost) Back to Statistics: Quasi-statistics: 1) delete one point from the dataset, 2) fitting, 3) analysis of the error for the deleted data; The overfitting problem and smoothed data points (it is very close to non- parametric statistics)

13 Principal manifolds Elastic maps framework SVM Principal manifolds Regression, approximation Supervised classification K- means SOM Clustering Multidim. scaling Visualization PCA Factor analysis LLE ISOMAP Non-linear Data-mining methods

14 Finite set of objects in R N X i i=1..m IRIS database Petal heght Petal width Sepal width Sepal height SPECIES Iris-setosa Iris-setosa Iris-setosa Iris-versicolor Iris-versicolor Iris-versicolor Iris-virginica X1.9Iris-virginica Iris-virginica Iris-virginica

15 Mean point K-means clustering

16 Principal Object,

17 Principal Component Analysis, Maximal dispersion 1 st Principal axis 2 nd principal axis

18 Principal manifold

19 Statistical Self-consistency x π π π -1 (x) x = E(y|π(y)=x) Principal Manifold

20 What do we want? Non-linear surface (1D, 2D, 3D …) Smooth and not twisted The data model is unknown Speed (time linear with Nm) Uniqueness Fast way to project datapoints

21 Metaphor of elasticity Data points Graph nodes U (Y) U (E), U (R)

22 Constructing elastic nets y E (0) E (1) R (1) R (0) R (2)

23 Definition of elastic energy. E (0) E (1) R (1) R (0) R (2) y XjXj

24 Elastic manifold

25 Global minimum and softening 0, , , ,

26 Adaptive algorithms Growing net Adaptive net Refining net: Idea of scaling:

27 Scaling Rules For uniform d-dimensional net from the condition of constant energy density we obtain: s is number of edges, r is number of ribs in a given volume

28 Grammars of Construction Substitution rules Examples: 1)For net refining: substitutions of columns and rows 2)For growing nets: substitutions of elementary cells.

29 Substitutions in factors × = Substitution rule Transformation of factor Graph factorization

30 Substitutions in factors × × Graph transformation

31 Transformation selection A grammar is a list of elementary graph transformations. Energetic criterion: we select and apply an elementary applicable transformation that provides the maximal energy decrease (after a fitting step). The number of operations for this selection should be in order O(N) or less, where N is the number of vertexes

32 Projection onto the manifold Closest node of the net Closest point of the manifold

33 Mapping distortions Two basic types of distortion: 1) Projecting distant points in the close ones (bad resolution) 2) Projecting close points in the distant ones (bad topology compliance)

34 Instability of projection Best Matching Unit (BMU) for a data point is the closest node of the graph, BMU2 is the second- close node. If BMU and BMU2 are not adjacent on the graph, then the data point is unstable. Gray polygons are the areas of instability. Numbers denote the degree of instability, how many nodes separate BMU from BMU2.

35 Colorings: visualize any function

36 Density visualization

37 Example: different topologies RNRN R2R2

38 VIDAExpert tool and elmap C++ package

39 Regression and principal manifolds regression principal component x F(x)

40 Projection and regression Data with gaps are modelled as affine manifolds, the nearest point on the manifold provides the optimal filling of gaps.

41 Iterative error mapping For a given elastic manifold and a datapoint x (i) the error vector is where P(x) is the projection of data point x (i) onto the manifold. The errors form a new dataset, and we can construct another map, getting regular model of errors. So we have the first map that models the data itself, the second map that models errors of the first model, … and so on. Every point x in the initial data space is modeled by the vector

42 Image skeletonization or clustering around curves

44 Approximation of molecular surfaces

45 Application: economical data Gross output Density Profit Growth temp

46 Medical table 1700 patients with infarctus myocarde Lethal cases Patients map, density

47 Medical table 1700 patients with infarctus myocarde 128 indicators Age Numberof infarctus in anamnesis Stenocardia functional class

48 Codon usage in all genes of one genome Escherichia coli Bacillus subtilis Majority of genes Highly expressed genes Foreign genes Hydrophobic genes

49 Golubs leukemia dataset 3051 genes, 38 samples (ALL/B-cell,ALL/T-cell,AML) ALL sample AML sample Map of genes: vote for ALL vote for AML used by T.Golub used by W.Lie

50 Golubs leukemia dataset map of samples: AML ALL/B-cell ALL/T-cell density Cystatin C Retinoblastoma binding protein P48 CA2 Carbonic anhydrase II X-linked Helicase II

51 Useful links Principal components and factor analysis Principal curves and surfaces Self Organizing Maps Elastic maps

52 Several names K-means clustering: MacQueen, 1967; SOM: T. Kohonen, 1981; Principal curves: T. Hastie and W. Stuetzle, 1989; Elastic maps: A. Gorban, A. Zinovyev, A. Rossiev, 1998; Polygonal models for principal curves: B. Kégl, 1999; Local PCA for orincipal curves construction: J. J. Verbeek, N. Vlassis, and B. Kröse, 2000.

53 Two of them are Authors

54 Thank you for your attention! Questions?

Еще похожие презентации в нашем архиве:

Готово:

Sequences Sequences are patterns. Each pattern or number in a sequence is called a term. The number at the start is called the first term. The term-to-term.

Sequences Sequences are patterns. Each pattern or number in a sequence is called a term. The number at the start is called the first term. The term-to-term.

© 2017 MyShared Inc.

All rights reserved.