<div dir="ltr">Hello Everyone,<br>I have uploaded my proposal on weighting scheme.<br><a href="https://www.google-melange.com/gsoc/proposal/review/student/google/gsoc2014/reetu11/5707702298738688">https://www.google-melange.com/gsoc/proposal/review/student/google/gsoc2014/reetu11/5707702298738688</a><br>
<div>Your comments will prove fruitful to me, so please kindly review it and let me know if any change is required.</div><div><br></div><div>Kind Regrads,<br>Reetesh</div></div><div class="gmail_extra"><br clear="all"><div>
<div dir="ltr">Reetesh Ranjan<br>Junior Undergraduate<div>IIT(BHU), Varanasi</div><div>contact no: +917275115929</div><div>Skype : reetesh.ranjan11</div></div></div>
<br><br><div class="gmail_quote">On Fri, Mar 21, 2014 at 5:30 PM, <span dir="ltr"><<a href="mailto:xapian-devel-request@lists.xapian.org" target="_blank">xapian-devel-request@lists.xapian.org</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Send Xapian-devel mailing list submissions to<br>
<a href="mailto:xapian-devel@lists.xapian.org">xapian-devel@lists.xapian.org</a><br>
<br>
To subscribe or unsubscribe via the World Wide Web, visit<br>
<a href="http://lists.xapian.org/mailman/listinfo/xapian-devel" target="_blank">http://lists.xapian.org/mailman/listinfo/xapian-devel</a><br>
or, via email, send a message with subject or body 'help' to<br>
<a href="mailto:xapian-devel-request@lists.xapian.org">xapian-devel-request@lists.xapian.org</a><br>
<br>
You can reach the person managing the list at<br>
<a href="mailto:xapian-devel-owner@lists.xapian.org">xapian-devel-owner@lists.xapian.org</a><br>
<br>
When replying, please edit your Subject line so it is more specific<br>
than "Re: Contents of Xapian-devel digest..."<br>
<br>
<br>
Today's Topics:<br>
<br>
1. GSoC 2014: Clustering of Search Results (Abhishek Gupta)<br>
2. Proposal update (Anish Kanchan)<br>
<br>
<br>
----------------------------------------------------------------------<br>
<br>
Message: 1<br>
Date: Fri, 21 Mar 2014 01:37:31 +0530<br>
From: Abhishek Gupta <<a href="mailto:a.gupps@gmail.com">a.gupps@gmail.com</a>><br>
To: <a href="mailto:xapian-devel@lists.xapian.org">xapian-devel@lists.xapian.org</a><br>
Subject: [Xapian-devel] GSoC 2014: Clustering of Search Results<br>
Message-ID:<br>
<CAPfSG7gVoO9Lj94V=bPTWceFeP40JiwBAQ2uZ=<a href="mailto:6-fXCKe4nJ%2BA@mail.gmail.com">6-fXCKe4nJ+A@mail.gmail.com</a>><br>
Content-Type: text/plain; charset="iso-8859-1"<br>
<br>
Sir,<br>
<br>
I am Abhishek Gupta. I know I am quite late for the project discussion<br>
because I came to know about GSoC a bit lately but still I would like to<br>
discuss this project which interests me a lot. I know I have to submit some<br>
code so as to show my skill set but as the deadline is quite near I will<br>
submit the patches or exercises after the deadline to strengthen my<br>
application and show my coding skill.<br>
I read your existing source code for the clustering which is quite slow<br>
because of the hierarchical based clustering which is not required at all.*You<br>
have already provided with the number of clusters you should have in the<br>
end*. So for this we can employ K-means algorithm which can perform far<br>
better than the current algorithm.<br>
<br>
1) Hierarchical clustering have high memory requirements *O(n*n)* in<br>
comparison to *O(n+K)* space complexity of K-means algorithm, where*n* is<br>
the number of elements and* K *is the number of clusters.<br>
2) Hierarchical clustering running time is *O(n*n*n)* on the other hand<br>
K-means algorithm has time complexity of *O(n)*.<br>
3) K-means improves the clustering iteratively, more you run the code more<br>
better you will get the results.<br>
<br>
One thing that K-means lacks is its non-deterministic outcome. Every time<br>
it will produce different clusters. But we can always run the algorithm<br>
10-12 times and then take the average even then it will perform far better<br>
than the hierarchical one.<br>
<br>
So I would like to propose this algorithm which can perform better than the<br>
hierarchical one. After that to improve the clustering more we can also<br>
implement K-medoids/K-means++ clustering methods.<br>
<br>
I would you give some reviews regarding the proposal, so that I can submit<br>
the proposal at time.<br>
<br>
Thanks and Regards<br>
Abhishek Gupta<br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <<a href="http://lists.xapian.org/pipermail/xapian-devel/attachments/20140321/7bda8301/attachment-0001.html" target="_blank">http://lists.xapian.org/pipermail/xapian-devel/attachments/20140321/7bda8301/attachment-0001.html</a>><br>
<br>
------------------------------<br>
<br>
Message: 2<br>
Date: Fri, 21 Mar 2014 09:59:06 +0530<br>
From: Anish Kanchan <<a href="mailto:anishgkanchan@gmail.com">anishgkanchan@gmail.com</a>><br>
To: Xapian Development <<a href="mailto:xapian-devel@lists.xapian.org">xapian-devel@lists.xapian.org</a>><br>
Subject: [Xapian-devel] Proposal update<br>
Message-ID:<br>
<CAJiBCaNXA77qcmcmciidGEGdXF6d=<a href="mailto:JvL3JK1ADiHzLtSjvALWg@mail.gmail.com">JvL3JK1ADiHzLtSjvALWg@mail.gmail.com</a>><br>
Content-Type: text/plain; charset="iso-8859-1"<br>
<br>
Hello,<br>
<br>
I have updated my proposal as per Olly's suggestion. Kindly review it and<br>
let me know if any changes are to be made.<br>
<br>
Thanks and Regards,<br>
Anish Kanchan<br>
Student, University of Mumbai<br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <<a href="http://lists.xapian.org/pipermail/xapian-devel/attachments/20140321/347d3b44/attachment-0001.html" target="_blank">http://lists.xapian.org/pipermail/xapian-devel/attachments/20140321/347d3b44/attachment-0001.html</a>><br>
<br>
------------------------------<br>
<br>
Subject: Digest Footer<br>
<br>
_______________________________________________<br>
Xapian-devel mailing list<br>
<a href="mailto:Xapian-devel@lists.xapian.org">Xapian-devel@lists.xapian.org</a><br>
<a href="http://lists.xapian.org/mailman/listinfo/xapian-devel" target="_blank">http://lists.xapian.org/mailman/listinfo/xapian-devel</a><br>
<br>
<br>
------------------------------<br>
<br>
End of Xapian-devel Digest, Vol 107, Issue 31<br>
*********************************************<br>
</blockquote></div><br></div>