Presentation is loading. Please wait.

Presentation is loading. Please wait.

軟體實作與計算實驗 1 Lecture 9 Classification Two mappings from position to color Linear combination of distances to centers RBF: Linear combination of exponential.

Published bySheena Campbell Modified over 9 years ago

Similar presentations

Presentation on theme: "軟體實作與計算實驗 1 Lecture 9 Classification Two mappings from position to color Linear combination of distances to centers RBF: Linear combination of exponential."— Presentation transcript:

1 軟體實作與計算實驗 1 Lecture 9 Classification Two mappings from position to color Linear combination of distances to centers RBF: Linear combination of exponential distances to centers Four computational steps (A) K-means (B) Cross Distances (C) Optimal Coefficients (D) Coloring Five flow charts: OC (optimal coefficients), Coloring, ER (error rate), TRAINING & TESTING :

2 軟體實作與計算實驗 2 Kmeans: Data and MATLAB codes data_9.zip demo_kmeans.m load data_9.mat plot(X(:,1),X(:,2),'.'); [cidx, Y] = kmeans(X,10); hold on; plot(Y(:,1),Y(:,2),'ro');

3 軟體實作與計算實驗 3 My_kmeans: Data and MATLAB codes data_9.zip demo_my_kmeans.m my_kmeans.m load data_9.mat plot(X(:,1),X(:,2),'.'); [Y] = my_kmeans(X,10); hold on; plot(Y(:,1),Y(:,2),'ro');

4 軟體實作與計算實驗 4 ClusteringTest.rar my_kmeans.m GUI

5 軟體實作與計算實驗 5

6 6 Path of 4 means to centers of clusters

7 軟體實作與計算實驗 7 Tracking Convergence of K means

8 軟體實作與計算實驗 8 Path to converge data_2c.zip

9 軟體實作與計算實驗 9 Color points A color point POSITION COLOR A mapping from position to color

10 軟體實作與計算實驗 10 Classification of color points A mapping from position to color A rule for discriminating red points from blue points

11 軟體實作與計算實驗 11 A mapping from position to color Each point has its membership to partitioned regions Blue points and red points are well separated A mathematical mapping NO TABLE LOOK-UP !!

12 軟體實作與計算實驗 12 load data_2c.mat data_2c.zip X=data_2c(:,1:2); Y=data_2c(:,3); X : positions of points Y : 0 or 1 ( blue or red color) Color points

13 軟體實作與計算實驗 13 Two-dimensional color points Load data for classification function show_color_data(X,Y) figure; ind = find(Y==0); plot(X(ind,1), X(ind,2), 'b. '); hold on ind = find(Y==1); plot(X(ind,1), X(ind,2), 'r.') load data_2c X=data_2c(:,1:2); Y=data_2c(:,3); show_color_data(X,Y)

14 軟體實作與計算實驗 14 Position & Color X and Y are paired data X collects positions of points Y collects colors of points A mapping from position to color is derived for classification of color points

15 軟體實作與計算實驗 15 Sampling for training Sampling one half as training points N=size(X,1); ind=randperm(N); n = floor(N/2); x_train=X(ind(1:n),:); y_train=Y(ind(1:n),:); show_color_data(x_train,y_train)

16 軟體實作與計算實驗 16 X and Y Sampling x_train and y_train Sampling for training n=N/2

17 軟體實作與計算實驗 17 x_train and y_train OC(optimal coefficients) STEPS A,B,C TRAINING STEP D: coloring x_trainy_haty_train ER (error rate) Error rate for training ½ data centers, a

18 軟體實作與計算實驗 18 X and Y TESTING STEP D: coloring XY_hatY ER (error rate) Error rate for training centers, a All data

19 軟體實作與計算實驗 19 Lower error rate for better testing Error Rate A mapping X All data

20 軟體實作與計算實驗 20 Linear cut load data_2c.txt error rate : zero

21 軟體實作與計算實驗 21 Linear cut x1x1 x2x2

22 軟體實作與計算實驗 22 Linear cut fails load data_3c.txt error rate : 0.289

23 軟體實作與計算實驗 23 Computations STEP A: kmeans Apply K-means to find K centers STEP B: cross distances Calculate cross distances between K means and N data points STEP C: optimal coefficients Determine a mapping from position to color

24 軟體實作與計算實驗 24 Sampling of training data data_3c.zip load data_3c X=data_3c(:,1:2); Y=data_3c(:,3); N=size(X,1); ind=randperm(N); n=floor(N/2); x_train=X(ind(1:n),:); y_train=Y(ind(1:n),:); show_color_data(x_train, y_train)

25 軟體實作與計算實驗 25 Step A K-means

26 軟體實作與計算實驗 26 centers = my_kmeans(x_train,3);

27 軟體實作與計算實驗 27 Step B Distances between centers and data points D(i,j) stores the distance between the ith point and the jth center D = cross_distance(x_train,centers) cross_distance.m

28 軟體實作與計算實驗 28 Step C: optimal coefficients Linear combination of distances to centers x … … a1a1 akak aKaK sum

29 軟體實作與計算實驗 29 A mapping for coloring (mapping I)

30 軟體實作與計算實驗 30 Coloring one point

31 軟體實作與計算實驗 31

32 軟體實作與計算實驗 32

33 軟體實作與計算實驗 33 STEP D: Coloring n points

34 軟體實作與計算實驗 34 STEP C: Optimal coefficients

35 軟體實作與計算實驗 35 Flow chart I : Optimal coefficients centers = my_kmeans(x_train,3 ); D = cross_distance(x_train,centers) x_train, y_train a = pinv(D)*y_train; centers, a STEP A STEP B STEP C funtion [centers,a]=OC(x_train,y_train)

36 軟體實作與計算實驗 36 x_train, centers, a y_hat=D*a; Flow chart II : coloring D = cross_distance(x_train,centers) STEP D coloring y_hat function y_hat=coloring(x_train,centers,a) STEP B

37 軟體實作與計算實驗 37 y_hat,y_train e=abs(y_train-round(y_hat)); n= length(y_train); er=sum(e)/n Error rate ans = 0 Flow chart III: Error rate for training function er=err_rate(y_hat,y_train)

38 軟體實作與計算實驗 38 Flow chart IV : TRAINING [centers,a]=OC(x_train,y_train) y_hat=coloring(x_train,centers,a) er=err_rate(y_hat,y_train) x_train,y_train er FLOW CHART I FLOW CHART II FLOW CHART III

39 軟體實作與計算實驗 39 Flow Chart V: TESTING load data_3c X=data_3c(:,1:2); Y=data_3c(:,3); Error rate ans = 0.026 y_hat=coloring(x_train,centers,a) er=err_rate(y_hat,y_train) centers,a FLOW CHART II FLOW CHART III

40 軟體實作與計算實驗 40 Mapping II from position to color x … … a1a1 akak aKaK sum Linear combination of exponential distances to centers

41 軟體實作與計算實驗 41 Discriminating rule

42 軟體實作與計算實驗 42

43 軟體實作與計算實驗 43 Set all variances to h for simplicity Step D: Coloring

44 軟體實作與計算實驗 44 Optimal coefficients

45 軟體實作與計算實驗 45 Flow chart I : Optimal coefficients centers = my_kmeans(x_train,3 ); D = cross_distance(x_train,centers) x_train, y_train h=200; a = pinv(exp(-D/h))*Y_TRAIN; centers, a STEP A STEP B STEP C funtion [centers,a]=OC(x_train,y_train)

46 軟體實作與計算實驗 46 x_train, centers, a y_hat=exp(-D/h)*a; Flow chart II : coloring D = cross_distance(x_train,centers) STEP D coloring Y_hat function y_hat=coloring(x_train,centers,a) STEP B

47 軟體實作與計算實驗 47 y_hat,y_train e=abs(y_train-round(y_hat)); n= length(y_train); er=sum(e)/n Error rate ans = 0 Flow chart III: Error rate for training function er=err_rate(y_hat,y_train)

48 軟體實作與計算實驗 48 Flow chart IV: TRAINING [centers,a]=OC(x_train,y_train) y_hat=coloring(x_train,centers,a) er=err_rate(y_hat,y_train) x_train,y_train er FLOW CHART I FLOW CHART II FLOW CHART III

49 軟體實作與計算實驗 49 Flow Chart V: TESTING load data_3c X=data_3c(:,1:2); Y=data_3c(:,3); y_hat=coloring(x_train,centers,a) er=err_rate(y_hat,y_train) Error rate ans = 0.000 centers,a FLOW CHART II FLOW CHART III

Download ppt "軟體實作與計算實驗 1 Lecture 9 Classification Two mappings from position to color Linear combination of distances to centers RBF: Linear combination of exponential."

Similar presentations

4.1- Plot Points in a Coordinate Plane

4.1- Plot Points in a Coordinate Plane
軟體實作與計算實驗 1  Roots of nonlinear functions  Nested FOR loops Lecture 5II.

軟體實作與計算實驗 1  Roots of nonlinear functions  Nested FOR loops Lecture 5II.
Component Analysis (Review)

Component Analysis (Review)
Exercise 1 In the ISwR data set alkfos, do a PCA of the placebo and Tamoxifen groups separately, then together. Plot the first two principal components.

Exercise 1 In the ISwR data set alkfos, do a PCA of the placebo and Tamoxifen groups separately, then together. Plot the first two principal components.
Types of Algorithms.

Types of Algorithms.
Computer vision: models, learning and inference Chapter 8 Regression.

Computer vision: models, learning and inference Chapter 8 Regression.
Radial Basis Functions

Radial Basis Functions
Coloring black boxes: visualization of neural network decisions Włodzisław Duch School of Computer Engineering, Nanyang Technological University, Singapore,

Coloring black boxes: visualization of neural network decisions Włodzisław Duch School of Computer Engineering, Nanyang Technological University, Singapore,
Data mining and statistical learning, lecture 5 Outline  Summary of regressions on correlated inputs  Ridge regression  PCR (principal components regression)

Data mining and statistical learning, lecture 5 Outline  Summary of regressions on correlated inputs  Ridge regression  PCR (principal components regression)
I.1 ii.2 iii.3 iv.4 1+1=. i.1 ii.2 iii.3 iv.4 1+1=

I.1 ii.2 iii.3 iv.4 1+1=. i.1 ii.2 iii.3 iv.4 1+1=
Prediction Networks Prediction –Predict f(t) based on values of f(t – 1), f(t – 2),… –Two NN models: feedforward and recurrent A simple example (section.

Prediction Networks Prediction –Predict f(t) based on values of f(t – 1), f(t – 2),… –Two NN models: feedforward and recurrent A simple example (section.
Bioinformatics Challenge  Learning in very high dimensions with very few samples  Acute leukemia dataset: 7129 # of gene vs. 72 samples  Colon cancer.

Bioinformatics Challenge  Learning in very high dimensions with very few samples  Acute leukemia dataset: 7129 # of gene vs. 72 samples  Colon cancer.
I.1 ii.2 iii.3 iv.4 1+1=. i.1 ii.2 iii.3 iv.4 1+1=

I.1 ii.2 iii.3 iv.4 1+1=. i.1 ii.2 iii.3 iv.4 1+1=
Basic Mathematics for Portfolio Management. Statistics Variables x, y, z Constants a, b Observations {x n, y n |n=1,…N} Mean.

Basic Mathematics for Portfolio Management. Statistics Variables x, y, z Constants a, b Observations {x n, y n |n=1,…N} Mean.
General Mining Issues a.j.m.m. (ton) weijters Overfitting Noise and Overfitting Quality of mined models (some figures are based on the ML-introduction.

General Mining Issues a.j.m.m. (ton) weijters Overfitting Noise and Overfitting Quality of mined models (some figures are based on the ML-introduction.
Discriminant Analysis Testing latent variables as predictors of groups.

Discriminant Analysis Testing latent variables as predictors of groups.
Radial Basis Function Networks

Radial Basis Function Networks
Ranga Rodrigo April 5, 2014 Most of the sides are from the Matlab tutorial. 1.

Ranga Rodrigo April 5, 2014 Most of the sides are from the Matlab tutorial. 1.
KE22 FINAL YEAR PROJECT PHASE 2 Modeling and Simulation of Milling Forces SIMTech Project Ryan Soon, Henry Woo, Yong Boon April 9, 2011 Confidential –

KE22 FINAL YEAR PROJECT PHASE 2 Modeling and Simulation of Milling Forces SIMTech Project Ryan Soon, Henry Woo, Yong Boon April 9, 2011 Confidential –
Chapter 4 Supervised learning: Multilayer Networks II.

Chapter 4 Supervised learning: Multilayer Networks II.

Similar presentations

Ads by Google