core points of dbscan clustering.

How to get the core points of each cluster in the dbscan. I need to store the core points in a variable. How can I do that? Please help me.

Answer 1

[Ameer Hamza]

Although this question is old, however, it is similar to your other question I answered. Therefore, If someone is searching for a solution to this question, please refer to my answer on a similar question:

https://www.mathworks.com/matlabcentral/answers/509704-number-of-elements-of-clusters-from-dbscan#answer_419114

[BHAGYALAKSHMI M]

Thank you sir. I tried this in my data set, but it giving some errors.

Error is --

Error using splitapply (line 61)

Group numbers must be a vector of positive integers, and cannot be a sparse vector.

Error in shpa (line 19)

centers = splitapply(@(x) mean(x, 1), core, core_idx);

How can I avoid the error and get the center. Please help me Sir.

clc;
clear;
data=xlsread('data.xlsx');
minpts=3;
epsilon=30;
[idx, corepts] = dbscan(data,epsilon,minpts);

fig1 = figure();
gscatter(data(:,1),data(:,2),idx);

fig2 = figure();
ax = axes();
hold on;

core=data(corepts, :);
core_idx = idx(corepts, :);
gscatter(core(:,1),core(:,2),core_idx);

centers = splitapply(@(x) mean(x, 1), core, core_idx);
gscatter(centers(:,1), centers(:,2), (1:6)');

[Ameer Hamza]

You modified the code a bit as compared to my previous answer. Please try the following code

clc;
clear;
data=readtable('data.xlsx', 'ReadVariableNames', false);
data.Var2 = findgroups(data.Var2); % convert column 
data.Var10(isnan(data.Var10)) = 0; % place 0 in empty cells
minpts=3;
epsilon=30;

data = table2array(data);
[idx, corepts] = dbscan(data,epsilon,minpts);

fig1 = figure();
gscatter(data(:,1),data(:,2),idx);

fig2 = figure();
ax = axes();
hold on;

core=data(corepts, :);
core_idx = idx(corepts, :);
gscatter(core(:,1),core(:,2),core_idx);

centers = splitapply(@(x) mean(x, 1), core, core_idx);
gscatter(centers(:,1), centers(:,2), 1:size(centers,1));

[BHAGYALAKSHMI M]

Thank you Sir.

The code is working. But if I change the datasets it giving the errors. How can I apply this to all datasets? Please help me.

[Ameer Hamza]

Please check the modified code below. Note that i removed the line since it was not necessary

data.Var10(isnan(data.Var10)) = 0; % place 0 in empty cells

Also you need to change the line according to each file. For example in file data2.xlsx there are two columns (column 1 and 9) that have text data. So you need to manually change the code to convert those columns to numbers

data.Var1 = findgroups(data.Var1); % convert column 
data.Var9 = findgroups(data.Var9); % convert column 

Change above codes according to each file.

clc;
clear;
data=readtable('data5.xlsx', 'ReadVariableNames', false);
data.Var5 = findgroups(data.Var5); % convert column 

minpts=3;
epsilon=30;

data = table2array(data);
[idx, corepts] = dbscan(data,epsilon,minpts);

fig1 = figure();
gscatter(data(:,1),data(:,2),idx);

fig2 = figure();
ax = axes();
hold on;

core=data(corepts, :);
core_idx = idx(corepts, :);
gscatter(core(:,1),core(:,2),core_idx);

centers = splitapply(@(x) mean(x, 1), core, core_idx);
gscatter(centers(:,1), centers(:,2), 1:size(centers,1));

[BHAGYALAKSHMI M]

Thank you Sir.