v1.1.9

nwequiv.ado

@@ -1,12 +1,12 @@
 capture program drop nwequiv
 program nwequiv
-	syntax [anything(name=netname)] [, type(string) generate(string) mode(string) *]
+	syntax [anything(name=netname)] [, type(string) iter(integer 3) generate(string) mode(string) *]
 
 	if "`generate'" == "" {
 		local generate "_eqvclass"
 	}
 
-	local name = "equivalent"
+	local name = "equivalence"
 	nwvalidate `name'
 	local name = r(validname)
 
@@ -63,27 +63,37 @@ program nwequiv
 		}
 		nwdrop __temp_geodesic
 	}
+	if ("`type'" == "regular"){
+		mata: S = regularEquivalence(onenet, `iter')
+		capture nwdrop _regEquiv
+		nwset, mat(S) name(`name')
+		mata: S
+		mata: S = (S:== 1)
+	}
 
-	nwset, mat(S) name(`name')
+	nwset, mat(S) name(__equiv)
 	capture drop `generate'
 	tempvar comp isol
 	qui nwgen `isol' = isolates(`netname')
-	qui nwcomponents `name', generate(`comp')
+	qui nwcomponents __equiv, generate(`comp')
 	qui replace `comp' = -1 if `isol' == 1
 	qui egen `generate' = group(`comp')
 	label variable `generate' "`generate'"
 	tab `generate', `options'
 
 	capture mata: mata drop S
 	capture mata: mata drop onenet
-	capture nwdrop `name'
+	capture nwdrop __equiv
 
 end
 
 
 capture mata: mata drop structuralEquivalence()
 capture mata: mata drop distanceEquivalence()
 capture mata: mata drop getDistribution()
+capture mata: mata drop getSimilarity()
+capture mata: mata drop getMatch()
+capture mata: mata drop regularEquivalence()
 
 mata:
 real matrix structuralEquivalence(real matrix net, scalar mode) {
@@ -167,5 +177,172 @@ real matrix getDistribution(real matrix X, scalar max){
 	}
 	return(result)
 }
+
+real matrix regularEquivalence(real matrix X, scalar iter) {
+	Sstart = J(rows(X), cols(X), 1)
+	for (i = 1 ; i <= iter; i++) {
+		Sstart = getSimilarity(X, Sstart)
+	}
+	return(Sstart)
+}
+
+mata:
+real matrix getSimilarity(real matrix X,real matrix S) {
+	S_new = S
+	for(i=1; i<= rows(X);i++){
+		for(j=1;j<= rows(X);j++){
+			S_new[i,j] = getMatch(X,S_new,i,j)
+		}
+	}
+	return(S_new)
+}
+
+
+real scalar getMatch(X,S,a,e) {
+	
+	points = 0
+	points_max = 0 
+	
+	nodes = rows(X)
+	temp = (1::nodes)
+	
+	a_outties = (editvalue((X[a,.] - X[.,a]'),-1,0))'
+	a_outties
+	a_out = select(temp, a_outties)
+	a_inties = (editvalue((X[.,a]' - X[a,.]),-1,0))'
+	a_in = select(temp, a_inties)
+	a_mutualties = (editvalue((X[a,.] + X[.,a]'),1,0))' :== 2
+	a_mutual = select(temp, a_mutualties)
+
+	e_outties = (editvalue((X[e,.] - X[.,e]'),-1,0))'
+	e_out = select(temp, e_outties)
+	e_inties = (editvalue((X[.,e]' - X[e,.]),-1,0))'
+	e_in = select(temp, e_inties)
+	e_mutualties = (editvalue((X[e,.] + X[.,e]'),1,0))' :== 2
+	e_mutual = select(temp, e_mutualties)	
+	
+	/////////
+	//	A's PERSPECTIVE
+	/////////
+	
+	// go through all out neighbors of a
+	for (i = 1 ; i <= rows(a_out); i++) {
+		bestmatch = 0
+		ii = a_out[i]
+		// look for a best match among e's neighbors and calculate points
+		for (j = 1; j<= rows(e_out); j++) {
+			jj = e_out[j]
+			if (S[ii,jj] > bestmatch) {
+				bestmatch = S[ii,jj]
+			}
+		}
+		points = points + bestmatch
+	}
+	// go through all in neighbors of a
+	for (i = 1 ; i <= rows(a_in); i++) {
+		ii = a_in[i]
+		bestmatch = 0
+		// look for a best match among e's neighbors and calculate points
+		for (j = 1; j<= rows(e_in); j++) {
+			jj = e_in[j]
+			if (S[ii,jj] > bestmatch) {
+				bestmatch = S[ii,jj]
+			}
+		}
+		points = points + bestmatch
+	}
+	// go through all mutual neighbors of a
+	for (i = 1 ; i <= rows(a_mutual); i++) {		
+		ii = a_mutual[i]
+		bestmatch = 0
+		// look for a best match among e's neighbors and calculate points
+		for (j = 1; j<= rows(e_mutual); j++) {
+			jj = e_mutual[j]
+			if (S[ii,jj] > bestmatch) {
+				bestmatch = 2 * S[ii,jj]
+			}
+		}
+		//WILDCARD find best match among e' in /out neighbors
+		if (bestmatch == 0) {
+			for (j = 1; j<= rows(e_in); j++) {
+				jj = e_in[j]
+				if (S[ii,jj] > bestmatch) {
+					bestmatch = S[ii,jj]
+				}
+			}
+			for (j = 1; j<= rows(e_out); j++) {
+				jj = e_out[j]
+				if (S[ii,jj] > bestmatch) {
+					bestmatch = S[ii,jj]
+				}
+			}
+		}
+		
+		points = points + bestmatch
+	}
+	
+	/////////
+	//	E's PERSPECTIE
+	/////////
+	
+	// go through all out neighbors of e
+	for (i = 1 ; i <= rows(e_out); i++) {
+		bestmatch = 0
+		ii = e_out[i]
+		// look for a best match among e's neighbors and calculate points
+		for (j = 1; j<= rows(a_out); j++) {
+			jj = a_out[j]
+			if (S[ii,jj] > bestmatch) {
+				bestmatch = S[ii,jj]
+			}
+		}
+		points = points + bestmatch
+	}
+	// go through all in neighbors of e
+	for (i = 1 ; i <= rows(e_in); i++) {
+		bestmatch = 0
+		ii = e_in[i]
+		// look for a best match among e's neighbors and calculate points
+		for (j = 1; j<= rows(a_in); j++) {
+			jj = a_in[j]
+			if (S[ii,jj] > bestmatch) {
+				bestmatch = S[ii,jj]
+			}
+		}
+		points = points + bestmatch
+	}
+	// go through all mutual neighbors of e
+	for (i = 1 ; i <= rows(e_mutual); i++) {
+		bestmatch = 0
+		ii = e_mutual[i]
+		// look for a best match among e's neighbors and calculate points
+		for (j = 1; j<= rows(a_mutual); j++) {
+			jj = a_mutual[j]
+			if (S[ii,jj] > bestmatch) {
+				bestmatch = 2 * S[ii,jj]
+			}
+		}
+		//WILDCARD find best match among a' in /out neighbors
+		if (bestmatch == 0) {
+			for (j = 1; j<= rows(a_in); j++) {
+				jj = a_in[j]
+				if (S[ii,jj] > bestmatch) {
+					bestmatch = S[ii,jj]
+				}
+			}
+			for (j = 1; j<= rows(a_out); j++) {
+				jj = a_out[j]
+				if (S[ii,jj] > bestmatch) {
+					bestmatch = S[ii,jj]
+				}
+			}
+		}
+		points = points + bestmatch
+	}
+
+	points_max = rows(a_out) + rows(a_in) + 2 * rows(a_mutual) + rows(e_out) + rows(e_in) + 2 * rows(e_mutual)
+	return(points /  points_max)
+}
 end
 
+

nwgeodesic.ado

@@ -138,9 +138,12 @@ real matrix distances(real matrix nw) {
 	found = nw
 	distance = nw
 	power = nw
-	
-	i= 1	
-	while (i<= (nodes / 2) & sum(found)< nodes*nodes - nodes ) {
+
+	num_found_previous = -1
+	num_found = 0
+	i = 1
+	while ((i<= (nodes / 2)) & (num_found < (nodes*nodes - nodes))) {
+	//while ((i<= (nodes / 2)) & (num_found < (nodes*nodes - nodes))  & (num_found > num_found_previous)) {
 		i = i+1
 		power = power * nw
 		power = power:/power
@@ -149,7 +152,9 @@ real matrix distances(real matrix nw) {
 		_editvalue(temp, -1, 0)
 		temp = temp * i
 		distance = distance + temp
+		num_found_previous = num_found
 		found = distance:/distance
+		num_found = sum(found)
 		_editmissing(found, 0)	
 	}
 	_editvalue(distance, 0, -1)

nwneighbor.ado

@@ -69,9 +69,14 @@ program nwneighbor
 	if temp_mat[1,1] == . {
 		local temp_rows = 0
 	}
+
+	local neighbors_list1 ""
+	local neighbors_list2 ""
 	forvalues j = 1/`temp_rows' {
 		local temp =  temp_mat[`j',1]
 		local onelab : word `temp' of `labs'
+		local neighbors_list1 "`neighbors_list1' `temp'"
+		local neighbors_list2 "`neighbors_list2' `onelab'"
 		if `uselab' == 1{
 			di "{res}`onelab'" _continue
 		}
@@ -82,6 +87,8 @@ program nwneighbor
 			di "{txt} , " _continue
 		}
 	}
+	mata: st_global("r(neighbors_list1)", "`neighbors_list1'")
+	mata: st_global("r(neighbors_list2)", "`neighbors_list2'")
 	di ""
 
 	di "{hline 40}"