tag:blogger.com,1999:blog-6314876008291942531.post656857333318820736..comments2024-01-14T00:36:43.430-08:00Comments on Antonio Gulli's coding playground: DBSCAN clustering algorithmUnknownnoreply@blogger.comBlogger10125tag:blogger.com,1999:blog-6314876008291942531.post-74200556259751855502013-01-30T13:26:54.000-08:002013-01-30T13:26:54.000-08:00Hi,
Thanks for sharing this code. The last com...Hi, <br /><br />Thanks for sharing this code. The last commenter (M^3 Team) 's correction is wrong. Similarity, not distance is being compared! I wrote this small extra bit of code to provide the example where one can see several clusters.<br /><br /> void randomInit_twopopulations (Points & ps, unsigned int dims,<br /> unsigned int num_points)<br /> {<br /> for (unsigned int j = 0; j < num_points; j++)<br /> {<br /> Point p(dims);<br /> <br /> double added;<br /> if (j < num_points/2)<br /> added = -5.0;<br /> else<br /> added = 5.0;<br /> <br /> for (unsigned int i = 0; i < dims; i++)<br /> {<br /> p(i) = (added + rand() * (2.0) / RAND_MAX);<br /> // std::cout << p(i) << ' ';<br /> }<br /> ps.push_back(p);<br /> // std::cout << std::endl;<br /> }<br /> }Unknownhttps://www.blogger.com/profile/01856875169703792111noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-13447882705173392602012-01-24T05:57:48.617-08:002012-01-24T05:57:48.617-08:00Hi
I've found a very very small bug in your c...Hi<br /><br />I've found a very very small bug in your code.<br />in clusters.cpp file<br /><br />in findNeighbours function:<br />if ((pid != j ) && (_sim(pid, j)) > threshold)<br /><br />it's supposed to be less than<br />if ((pid != j ) && (_sim(pid, j)) < threshold)<br /><br />after changing this it works like charm, thanks for the codeM^3 Teamhttps://www.blogger.com/profile/05065702838558886689noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-39779894596343698042012-01-02T07:51:02.525-08:002012-01-02T07:51:02.525-08:00thanks for sharing the code..where I can find real...thanks for sharing the code..where I can find real dataset for applying DBSCAN?saharhttps://www.blogger.com/profile/00508328172426708869noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-14964719326921915532011-05-27T05:53:47.905-07:002011-05-27T05:53:47.905-07:00actually I tried this code on ELKI example, it pro...actually I tried this code on ELKI example, it produced just one cluster, where ELKI gave several clusters with the same parameters.Jaravinehttps://www.blogger.com/profile/04149811400848881095noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-78711624037517516022011-05-27T04:56:23.852-07:002011-05-27T04:56:23.852-07:00The example is clustering random numbers into 1 cl...The example is clustering random numbers into 1 cluster. Actually would be good if the example was in some way illustrative of this method - on a more "shapy" data - like in the paper on DBSCAN.Jaravinehttps://www.blogger.com/profile/04149811400848881095noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-20770626350049327822010-03-03T06:40:22.975-08:002010-03-03T06:40:22.975-08:00At lines 16 and 33 in distance.h I had to change
...At lines 16 and 33 in distance.h I had to change<br /><br />typedef typename VEC_T vector_type;<br />to<br />typedef VEC_T vector_type;<br /><br />to get it to compile.<br /><br />DaveDavid Doriahttps://www.blogger.com/profile/01804778285428434637noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-79542080523861978782010-02-06T15:22:49.957-08:002010-02-06T15:22:49.957-08:00hi,
Thank you for sharing the codes. However, I ...hi, <br /><br />Thank you for sharing the codes. However, I met a problem when I tried to compile the codes in visual studio. Then, I installed boost on my computer. However, it does not work out. Could you please show me some possible solutions to this problem? <br /><br />Thank you so much!Unknownhttps://www.blogger.com/profile/02881859196348025058noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-24246583795765776422010-02-06T15:19:59.585-08:002010-02-06T15:19:59.585-08:00Hi,
Thank you for sharing the codes. However, I m...Hi, <br />Thank you for sharing the codes. However, I met a problem when I try to compile the codes with visual studio 2005. I had installed boost on my computer, but it still does not work out. Can you please show me some possible solutions to this problem? <br /><br />Thank you so much!Unknownhttps://www.blogger.com/profile/02881859196348025058noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-55027280639286971552009-11-27T02:50:36.690-08:002009-11-27T02:50:36.690-08:00Thanks for releasing the code. However; could you ...Thanks for releasing the code. However; could you specify the license under which it is released?<br /><br />ThanksUnknownhttps://www.blogger.com/profile/18323256628611745438noreply@blogger.comtag:blogger.com,1999:blog-6314876008291942531.post-74146445189351301932009-11-17T04:09:04.480-08:002009-11-17T04:09:04.480-08:00Thanks, for sharing the code.
That would be nice t...Thanks, for sharing the code.<br />That would be nice to see how the DBSCAN will work if one would use the local Adaptive Metric(Mahalanobis metric) for the distance instead of Euclidean distance.<br />I am trying to implement it for 6D(x,y,z,vx,vy,vz)<br /><br />thanks <br />Arman.Anonymoushttps://www.blogger.com/profile/02706805224856951625noreply@blogger.com