Comments on: Cheap GPS and Code Project Tutorial

By: Shobhit Paliwal

Shobhit Paliwal — Tue, 19 Apr 2011 13:54:19 +0000

To Eric,
The clustering doesn’t seems to be work due to an error caused by imatrix maths operation…an unidentified assembly or reference not found being pointed out …

please help me out in this one..

By: vijay

vijay — Wed, 23 Feb 2011 09:31:36 +0000

Dear Eric,

I had a couple of problems with ur beautiful kmeans doc clus.

1. I hav Dundas Chart professional edition, which didnt support the following statements:
i. DataSet myDocs = Chart1.DataManipulator.ExportSeriesValues(“Series1″);
ii. DataSet myKMeansPoints = Chart1.DataManipulator.ExportSeriesValues(“Series3″);

So, I tried as follows:
DataSet myDocs = new DataSet();
DataTable dt1 = new DataTable(“Series1″);
myDocs.Tables.Add(dt1);
DataPoint temppt = Chart1.Series["Series1"].Points[0];
dt1.Columns.Add(“X”);
dt1.Columns.Add(“Y”);
for(int c=2; c dt1.Columns.Add("Y"+c);
for (int j=0;j {
DataPoint p = Chart1.Series["Series1"].Points[j];
DataRow row = dt1.NewRow();
row[0] = p.XValue;
for (int i = 0; i < p.YValues.Length; i++)
row[i+1] = p.YValues[i];
dt1.Rows.Add(row);
}
for both Series1[mydocs] and Series3[mykmeanspoints]

- now that compiled and gave me some gud results. Can u please evaluate my code.

2. after the above changes, if i remove the comments for
//double[,] myWords = mylsi.MyWords;
//PlotWords(myDocs, mylsi.MyWordsRowCount);
am getting indexoutofrange problem.
i observed that the matrix has 12 rows and 9 cols.
The problem is in PlotWords method:
private void PlotWords(double[,]myWords,int myWordsRowCount)
{
for (int i = 0; i size 9 X 2
// myWordsRowCount -> 12
// “Matrix words = reducedWordVector*reducedSigma” ???
}
…

Can u please help me out here !

By: Eric

Eric — Sat, 17 Apr 2010 05:18:45 +0000

Hey Salma - The k-means class on http://eric.ness.net/archives/k-means-document-clustering/ is in C#. Or over at http://www.codeproject.com/KB/recipes/K-Mean_Clustering.aspx. Let me know if you have any other questions.

By: salma

salma — Sat, 17 Apr 2010 05:02:11 +0000

hi i need source code for k-means clustering algorithm for documents…in c#.plz any one have this code then send to my mail.Thanks in advance.

By: Aresh

Aresh — Tue, 17 Nov 2009 11:47:09 +0000

interesting implementation!

By: Radski

Radski — Tue, 01 Sep 2009 22:45:15 +0000

Thanks very much, same book same error…

By: Nadya

Nadya — Wed, 05 Aug 2009 12:30:20 +0000

Eric,

thank you very much for your answer. I spent one day trying to understand why my calculattions for Toy and LaSalle in the book do not match values i get.

Soriano,
Thank you very much for raising this issue so readers can find an answer on this page.

Nadya

By: Eric

Eric — Sun, 07 Jun 2009 04:26:49 +0000

Hey Soriano,

I think you are correct in pointing out this error. My guess is that probably got missed during editing.

Wikipedia has it stated as thus:

Two-dimensional distance

For two 2D points, P=(p_x,p_y)\, and Q=(q_x,q_y)\,, the distance is computed as:

\sqrt{(p_x-q_x)^2 + (p_y-q_y)^2}.

http://en.wikipedia.org/wiki/Euclidean_distance

So in python it should read as you’ve written it:

>> sqrt(pow(1.0-2.0,2)+pow(4.5-4.0,2))
1.11803398875

Good catch!

Eric

By: Soriano

Soriano — Sun, 07 Jun 2009 02:03:43 +0000

Hi,
IÂ´m reading the same book, and I had an doubt.
In “Euclidean Distance Score” the autor have been used
a simple function:

>> sqrt(pow(5-4,2)+pow(4-1,2))
3.1622776601683795

The values used in the function, the author
have been talked to calculate the distance
between Toby and LaSalle in the chart on the
figure 2-1.

But in the chart Toby has Snakes 4.5 and Dupree to 1.0
and LaSalle 4.0 to Snakes and 2.0 to Dupree.

My question is:
Why he didnÂ´t used this values as below.

D(Toby,LaSalle) =
>> sqrt(pow(1.0-2.0,2)+pow(4.5-4.0,2))
3.1622776601683795

Regards,

Soriano from Brazil