Initial commit

.gitattributes

@@ -0,0 +1,2 @@
+# Auto detect text files and perform LF normalization
+* text=auto

ACC.m

@@ -0,0 +1,6 @@
+function result = ACC(Y, predY)
+%if pred_classnum
+res = bestMap(Y, predY);
+% accuarcy
+result = length(find(Y == res))/length(Y);
+end

Assign_Control.m

@@ -0,0 +1,46 @@
+function pv_point = Assign_Control(sList,X,k,pi)
+% Calculate the percentage of samples controlled by the assignment
+%% assignment: pval(thisK)<pval(otherK)
+OtherClusters = setdiff(unique(pi),k);
+numClu = size(sList,1);
+numOther = length(OtherClusters);
+pv_point = zeros(numClu,numOther);
+for kth = 1:length(OtherClusters)
+otherK = find(pi==OtherClusters(kth));
+OtherS = X(otherK,:);
+% X_s = X(otherK,:);
+pvs = Other_single_point_fisher_exactG(X, sList, OtherS);
+for sth=1:numClu
+    pv_point(sth,kth) = binomtest(pvs(sth,:),0.05);
+end
+end
+end
+
+function p = Other_single_point_fisher_exactG(X, sList, OtherS)
+% He Z, Zhao C, Liang H, et al. Protein complexes identification with family-wise error rate control[J].
+% IEEE/ACM transactions on computational biology and bioinformatics, 2019, 17(6): 2062-2073.
+% Lancichinetti A, Radicchi F, Ramasco J J, et al. Finding statistically significant communities in networks[J]. 
+% PloS one, 2011, 6(4): e18961.
+% record of revisions:
+%     date               programmer              description of change
+% -----------        -----------------          ------------------------
+%  Nov 20, 2023           Lianyu Hu              Original code version
+%% for each attribute value u of other cluster OtherS, compute p-value on point sList
+[N,num_attr] = size(X);
+i_S = size(sList,1);
+i_X = X(sList(:,1),:);
+D_S = size(OtherS,1);
+p = zeros(i_S,num_attr);
+for attr=1:num_attr
+    X_attr = X(:,attr);
+    [value, tab] = histRate(X_attr);
+    for th = 1:i_S
+        s = i_X(th,attr);
+        k_i = tab(value==s);
+        k_in =  sum(OtherS(:,attr)==s);
+        k_out = k_i - k_in;
+        volume = min(D_S-k_in, k_out);
+        p(th,attr) = sum(hygepdf(k_in:k_in+volume,N,D_S,k_i));
+    end
+end
+end

CMI.m

@@ -0,0 +1,12 @@
+function result = CMI(X,pi)
+% Cluster membership index (CMI)
+N = size(X,1);
+K = max(pi);
+signum = zeros(K,1);
+parfor k =1:K
+    clusterk = find(pi==k);
+    pval = SigCM_intra(X,clusterk);
+    [signum(k,1),~,~] = FWER_Control(pval);
+end
+result = sum(signum)/N;
+end

Datasets/balance-scale.names

@@ -0,0 +1,57 @@
+1. Title: Balance Scale Weight & Distance Database
+
+2. Source Information: 
+    (a) Source: Generated to model psychological experiments reported
+		by Siegler, R. S. (1976).  Three Aspects of Cognitive
+		Development.  Cognitive Psychology, 8, 481-520.
+    (b) Donor: Tim Hume ([email protected])
+    (c) Date: 22 April 1994
+
+3. Past Usage: (possibly different formats of this data)
+   - Publications
+	1. Klahr, D., & Siegler, R.S. (1978).  The Representation of
+	   Children's Knowledge.  In H. W. Reese & L. P. Lipsitt (Eds.),
+	   Advances in Child Development and Behavior, pp. 61-116.  New
+	   York: Academic Press 
+	2. Langley,P. (1987).  A General Theory of Discrimination
+	   Learning.  In D. Klahr, P. Langley, & R. Neches (Eds.),
+	   Production System Models of Learning and Development, pp.
+	   99-161. Cambridge, MA: MIT Press
+	3. Newell, A. (1990).  Unified Theories of Cognition.
+	   Cambridge, MA: Harvard University Press
+	4. McClelland, J.L. (1988).  Parallel Distibuted Processing:
+	   Implications for Cognition and Development.  Technical
+	   Report AIP-47, Department of Psychology, Carnegie-Mellon
+	   University 
+	5. Shultz, T., Mareschal, D., & Schmidt, W. (1994).  Modeling
+	   Cognitive Development on Balance Scale Phenomena. Machine
+	   Learning, Vol. 16, pp. 59-88.
+
+4. Relevant Information: 
+	This data set was generated to model psychological
+	experimental results.  Each example is classified as having the
+	balance scale tip to the right, tip to the left, or be
+	balanced.  The attributes are the left weight, the left
+	distance, the right weight, and the right distance.  The
+	correct way to find the class is the greater of 
+	(left-distance * left-weight) and (right-distance *
+	right-weight).  If they are equal, it is balanced.
+
+5. Number of Instances: 625 (49 balanced, 288 left, 288 right)
+
+6. Number of Attributes: 4 (numeric) + class name = 5
+
+7. Attribute Information:
+	1. Class Name: 3 (L, B, R)
+	2. Left-Weight: 5 (1, 2, 3, 4, 5)
+	3. Left-Distance: 5 (1, 2, 3, 4, 5)
+	4. Right-Weight: 5 (1, 2, 3, 4, 5)
+	5. Right-Distance: 5 (1, 2, 3, 4, 5)
+
+8. Missing Attribute Values: 
+	none
+
+9. Class Distribution: 
+   1. 46.08 percent are L
+   2. 07.84 percent are B
+   3. 46.08 percent are R