svmplugin Discussions Rss Feed

New Post: Installation Experience SVM 1.0

jonathanswalton — Wed, 16 May 2012 09:53:22 GMT

We have tried to install the plug-in on an XP box running 2005. The msi file runs to completion and reports no errrors, but the algorithm isn't then visible via Visual Studio or Management Studio (across user and admin accounts). Installing components manually, again no errors, but the algorithm isn't visible.

Any suggestions?

New Post: Installation Experience SVM 1.0

MarkTab — Mon, 20 Sep 2010 20:28:27 GMT

Hi Furmangg

Hopefully someone will post an x64 solution soon.

I rewrote the first manual installation file in PowerShell script for Windows 7 x64 (though these commands will not succeed without a new build targeting x64):

& "C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\Bin\gacutil.exe" /u DMPluginWrapper
& "C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\Bin\gacutil.exe" /if DMPluginWrapper.dll
& "C:\Windows\Microsoft.NET\Framework\v2.0.50727\RegAsm.exe" SupportVectorMachine.dll
& "C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\Bin\gacutil.exe" /u SupportVectorMachine
& "C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\Bin\gacutil.exe" /if SupportVectorMachine.dll
& "C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\Bin\gacutil.exe" /u SMO
& "C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\Bin\gacutil.exe" /if SMO.dll

For the CMD prompt, the variable %ProgramFiles(x86)% would be the correct stem environment variable, but as you see in my code, the location of gacutil.exe is different for my installation of VS2010 (compared with "Microsoft Visual Studio 8\SDK\v2.0\Bin\gacutil.exe").

For PowerShell, the third file would be:

stop-service mssqlserverolapservice
start-service mssqlserverolapservice

This third file is the same as doing a "Restart" from right-clicking the Analysis Services instance in SSMS.

I was starting to write a generic PowerShell installation, but I have no way to test it, and I do NOT know all the possible combinations of Windows OS which might need to be handled. And, I'm understanding that the program needs to be recompiled, but again, I'm not aware of all the possible paths installation can take, so I am NOT the right person to provide a generic installation. [http://social.msdn.microsoft.com/Forums/en-US/sqldatamining/thread/76aaea48-b7cb-4171-a156-22aa66d138af/]

I am writing about this subject today because blogged about this topic last Saturday, and decided to follow up again on the SVM: http://www.marktab.net/datamining/index.php/2010/09/18/extending-sql-server-data-mining/

New Post: Installation Experience SVM 1.0

furmangg — Fri, 10 Sep 2010 23:07:25 GMT

Anybody got a working 64-bit installer built they can upload?

New Post: Installation Experience SVM 1.0

jorisvalkonet — Wed, 25 Feb 2009 19:47:49 GMT

Thanks for using the plug-in.

There is a problem with X64 versions. [http://social.msdn.microsoft.com/Forums/en-US/sqldatamining/thread/76aaea48-b7cb-4171-a156-22aa66d138af/] We have to compile a part of the plugin in X64 and post a X64 version of the algorithm. Apologies for your trouble.

We are currently working on a demo with the Adventure Works data and I understand the problems you encountered. We will try to fix this in the next release. If you have adjusted code which you like to see in the next version, then please send me the code. We are working on the next version besides working at customers, so the development of the next version is not as rapid as I would like..

Thanks for your feedback. It is very helpful.

Joris

New Post: Installation Experience SVM 1.0

marktab — Wed, 25 Feb 2009 04:13:51 GMT

Thanks for working on this plug-in for this important data mining algorithm.

I was able to install the plug-in and viewer on two separate machines running Vista 32 SP1 with SQL Server 2008. The plug-in (on both machines) shows in Excel (with the Data Mining plug-in) and in BIDS too (so does the viewer).

I was not able to install the plug-in and viewer on a Windows Server 2008 64-bit with SQL Server 2008. I therefore followed the manual installation instructions, and was able to modify the BAT file to point to the "(x86)" directory (as appopriate) which is separate from "Program Files" (for 64 bit). I used the command lines individually in a command window with Administrative privileges, and was able to interactively see that each DLL was registered. Using SSMS, the XMLA script shows that it did work. When I examined a copy of the INI file (I was not able to open the regular INI file, some sort of security setting) I could see the Avanade code in the XML (adding the option to the algorithm list). However, the plug-in does not show on either Excel or BIDS on this instance.

To test the algorithm, I added it to the "Targeted Mailing" structure in the Adventure Works demo. On one Vista machine, the project deployed, and the model processed. When I went to view the model, there was some error because drillthrough was not enabled for the model (which is the default), so it crashed VS 2008. When I restarted VS 2008, I reprocessed the model with drillthrough on. However, I am not finding a combination of x-axis and y-axis variables which is showing anything on the display. All I get is the spinning circle when I press "Load" (which is what I assume I should be doing since the "Load" button is in red). Red is normally a color indicating emergency or crisis, rather than process, so a simpler interface would simply be larger text or bold -- that's enough to distinguish it. Also, it would be better to say "Process" or "Display" or some similar word because "Load" has a specific meaning with data processing, and though the viewer is being loaded, we are not really loading data into some warehouse or UDM.

I can see the viewer code and I could make the changes to the load button. However, I'm hoping for something more important: a run that works.

By the way, it finally finished with a standard timeout error:
XML for Analysis parser: The XML for Analysis request timed out before it was completed.
Internal error: An unexpected exception occurred.
COM error: COM error: DMPlugInWrapper; Object reference not set to an instance of an object..
XML for Analysis parser: The XML for Analysis request timed out before it was completed.

One simple fix right now would be to add a project to the download area which assumes the Adventure Works data. This sample project would have all the parameters set correctly, and once processed, would allow someone to fully test their installation of the plug-in and viewer.

New Post: Functionality for the next version

jorisvalkonet — Mon, 19 Jan 2009 08:18:15 GMT

Currently, we are looking for an intern to develop the regression version of SVM. So, the next version will probably host the regression version of SVMs.

Joris Valkonet

New Post: Functionality for the next version

lukadt — Sat, 17 Jan 2009 15:53:19 GMT

I'm happy you will implement grid search parameter, with the new cross validation added to sql server 2008 it should be quite easy...
i'd like to know if you plan to extend your implementation to include also svm regression or do you think that actual data mining algorithms already cover regression? To conclude, a clear and detailed tutorial would be really appreciated by users
thanks in advance,
Luca Del Tongo

New Post: Functionality for the next version

jorisvalkonet — Sun, 11 Jan 2009 12:51:01 GMT

Thank you for your feedback. I agree that the grid search parameter selection would be beneficial to the plug-in.

New Post: Functionality for the next version

lukadt — Sat, 10 Jan 2009 22:49:37 GMT

You have done agood job, i think that you should add the grid search parameter selection (c and gamma) for achieving better result...
as you know the performance of an svm heavily depends on these parameters...
Hope this helps,
Luca Del Tongo

New Post: Tutorial?

lukadt — Sat, 10 Jan 2009 22:45:52 GMT

Just a little bit correction: increasing the value of C increases the cost of misclassifying points and forces the creation of a more accurate model that may not generalize well (overfitting)... keep up the good work,
Luca Del Tongo

New Post: Tutorial?

JorisValkonet — Tue, 11 Nov 2008 20:16:13 GMT

We didn't implement anything for the probability of a prediction. We might look into this and try to release that in the next version of the plug-in.

Joris

New Post: Tutorial?

pauloebc — Mon, 10 Nov 2008 16:58:00 GMT

Tks a lot! It works!

I just couldn't see the viewer working. But thats ok.

So finally, is there any probablity associated to predicted?

Best!!!

Paulo

On Mon, Nov 10, 2008 at 12:04 PM, JorisValkonet <notifications@codeplex.com> wrote:

From: JorisValkonet

We implemented the default 'Predict' for prediction. You should be able to run the following query:

SELECT m.[predict_column] as [prediction], x.[predict_column] as [input], x.*
FROM [model] As m
NATURAL PREDICTION JOIN
(SELECT * from [model].CASES)
AS x

If you alter the query and replace predict_column with the predict column of your model and [model] with your model, then you should be able to see the prediction.

The methods: "GetAttributes", "GetClasses", "GetPredictAttribute" and "GetMaxMinValues" are used for the viewer. These methods don't provide prediction functionality. (I think that I removed the visibility of the predict function; it is there and you can use it, but you are unable to see it).

Read the full discussion online.

To add a post to this discussion, reply to this email (svmplugin@discussions.codeplex.com)

To start a new discussion for this project, email svmplugin@discussions.codeplex.com

You are receiving this email because you subscribed to this discussion on CodePlex. You can unsubscribe on codePlex.com.

Please note: Images and attachments will be removed from emails. Any posts to this discussion will also be available online at codeplex.com

--
Abraços,
_______________________________
Paulo Carvalho
when you think that you know all answers, life comes and change all questions...
brilliant minds discuss ideas, half minds discuss events, dumb minds discuss people...

New Post: Tutorial?

JorisValkonet — Mon, 10 Nov 2008 14:04:46 GMT

We implemented the default 'Predict' for prediction. You should be able to run the following query:

SELECT m.[predict_column] as [prediction], x.[predict_column] as [input], x.*
FROM [model] As m
NATURAL PREDICTION JOIN
(SELECT * from [model].CASES)
AS x

If you alter the query and replace predict_column with the predict column of your model and [model] with your model, then you should be able to see the prediction.

The methods: "GetAttributes", "GetClasses", "GetPredictAttribute" and "GetMaxMinValues" are used for the viewer. These methods don't provide prediction functionality. (I think that I removed the visibility of the predict function; it is there and you can use it, but you are unable to see it).

New Post: Tutorial?

pauloebc — Mon, 10 Nov 2008 13:50:30 GMT

Tks to clarify some stuffs about the algorithm...

Look me here again...

I've ran a simple model

Predict: 0/1

Input:

Car = 0/1

Retired = 0/1

In a sample with 1000 rows. Using default settings, just changing to RBF.

and then I got the error attached...

About the results, when I use neural networks, I used to use the Preditc Fuction to return, for the followings Securit ID's, its predicted results (0/1 = good/bad).

But on SVM I just found "GetAttributes", "GetClasses", "GetPredictAttribute" and "GetMaxMinValues", but neither one have returned me the predicted. Actually neither one have returned anything.

I'm really sorry about all these questions, but note that I want to solve this issue using your tool, b/c sounds me very trustable.

Best

Paulo

On Mon, Nov 10, 2008 at 10:52 AM, JorisValkonet <notifications@codeplex.com> wrote:

From: JorisValkonet

Hi Paulo,

Thank you for your reply. I believe that SVMs could be the solution for your problem. I will try to help you.

1. Could give a hand telling how to configure the algorithm to predict a variable good/bad from a list of inputs?

Before running the algorithm, you have to do two things:

Create a data mining structure describing the meta data of your dataset. For example, you specify which columns are used and which column is the key column. Furthermore, you specify the cross-validation settings with the data mining structure (SQL 2008 only).

After you created your data mining structure, you have to create a data mining model. In the data mining model you have to specify which columns are input and what the predict column is. The input columns can be continuous and discrete. The predict column can only be discrete. Note that in this first version, you have to specify a key columns and nested tables (semi multi-relational data mining) is not supported. Besides specifying the types of columns, you can specify the kernel with the algorithm parameters. I believe that with your problem a linear or RBF (gaussian) kernel would be the ones I would pick.

These two steps can be done from Excel with the MSFT Office 2007 Excel Data Mining add-in or from Business Intelligence Development Studio or even from code in SQL Server Management Studio. The language DMX can be used for all data mining related tasks. Here you can find the language reference: http://technet.microsoft.com/en-us/library/ms132058.aspx. It might be fun to watch the videos of Rafal Lukawiecki about data mining: http://www.microsoft.com/emea/spotlight/event.aspx?id=99 (or use a search engine to find the videos). This should provide you a solid base for development.

2. How many rows approximally supports the algorithm?

The performance of the algorithm depends on 3 things:

The complexity of the problem in relation to the ability of the kernel to solve the problem

The number of rows

The number of dimensions

The first point can be tackeled by choosing the right kernel with the right parameters. I believe this is the most difficult part of SVMs; there are numerous papers out there how to solve this. Personally, I use a trail and error approach.
There is no fixed number of rows the algorithm takes. In theory the number of rows can be very large. I believe that it's safe to offer about 2000 rows. The performance is not only depending on the number of rows, but also of the dimensionality of the problem.
The dimensionality depends on the number of columns and the type of the columns. SVMs work in continuous space and therefore a continuous input column is a dimension for the algorithm. A discrete column is transformed to binairy continuous input space. The number of dimensions a discrete column has depends on the number of different values the column has. Each value translates into an additional dimension. For example if a dataset has the column 'X' with the potential values {'A','B','C'}. If the row with Id 1 has the value 'B ' for column X, then this is translated to:

Id X.A X.B X.C

1 0 1 0

So, the more discrete columns with discrete values there are, the more dimensions the SVM problem has. This can grow very fast and the running time is really influenced by this. The influence of continous input columns is relatively small.

3. How many inputs supports (how many dimentions)?
There is no limit, but you should consider the text above. with each discrete input column, the dimensionality grows fast.

4. Whats the fuction do I need to use to put the predict in a table?

You can use the Excel plug-in for predictions or you can use code. How to write DMX code for predictions: http://technet.microsoft.com/en-us/library/ms132031.aspx

Thanks for the feedback about the viewer.

If you have more/ additional questions, let me know.

Regards,
Joris

Read the full discussion online.

To add a post to this discussion, reply to this email (svmplugin@discussions.codeplex.com)

To start a new discussion for this project, email svmplugin@discussions.codeplex.com

You are receiving this email because you subscribed to this discussion on CodePlex. You can unsubscribe on codePlex.com.

Please note: Images and attachments will be removed from emails. Any posts to this discussion will also be available online at codeplex.com

New Post: Tutorial?

JorisValkonet — Mon, 10 Nov 2008 12:51:58 GMT

Hi Paulo,

Thank you for your reply. I believe that SVMs could be the solution for your problem. I will try to help you.

1. Could give a hand telling how to configure the algorithm to predict a variable good/bad from a list of inputs?

Before running the algorithm, you have to do two things:

Create a data mining structure describing the meta data of your dataset. For example, you specify which columns are used and which column is the key column. Furthermore, you specify the cross-validation settings with the data mining structure (SQL 2008 only).
After you created your data mining structure, you have to create a data mining model. In the data mining model you have to specify which columns are input and what the predict column is. The input columns can be continuous and discrete. The predict column can only be discrete. Note that in this first version, you have to specify a key columns and nested tables (semi multi-relational data mining) is not supported. Besides specifying the types of columns, you can specify the kernel with the algorithm parameters. I believe that with your problem a linear or RBF (gaussian) kernel would be the ones I would pick.

These two steps can be done from Excel with the MSFT Office 2007 Excel Data Mining add-in or from Business Intelligence Development Studio or even from code in SQL Server Management Studio. The language DMX can be used for all data mining related tasks. Here you can find the language reference: http://technet.microsoft.com/en-us/library/ms132058.aspx. It might be fun to watch the videos of Rafal Lukawiecki about data mining: http://www.microsoft.com/emea/spotlight/event.aspx?id=99 (or use a search engine to find the videos). This should provide you a solid base for development.

2. How many rows approximally supports the algorithm?

The performance of the algorithm depends on 3 things:

The complexity of the problem in relation to the ability of the kernel to solve the problem
The number of rows
The number of dimensions

The first point can be tackeled by choosing the right kernel with the right parameters. I believe this is the most difficult part of SVMs; there are numerous papers out there how to solve this. Personally, I use a trail and error approach.
There is no fixed number of rows the algorithm takes. In theory the number of rows can be very large. I believe that it's safe to offer about 2000 rows. The performance is not only depending on the number of rows, but also of the dimensionality of the problem.
The dimensionality depends on the number of columns and the type of the columns. SVMs work in continuous space and therefore a continuous input column is a dimension for the algorithm. A discrete column is transformed to binairy continuous input space. The number of dimensions a discrete column has depends on the number of different values the column has. Each value translates into an additional dimension. For example if a dataset has the column 'X' with the potential values {'A','B','C'}. If the row with Id 1 has the value 'B ' for column X, then this is translated to:

Id	X.A	X.B	X.C
1	0	1	0

So, the more discrete columns with discrete values there are, the more dimensions the SVM problem has. This can grow very fast and the running time is really influenced by this. The influence of continous input columns is relatively small.

3. How many inputs supports (how many dimentions)?
There is no limit, but you should consider the text above. with each discrete input column, the dimensionality grows fast.

4. Whats the fuction do I need to use to put the predict in a table?

You can use the Excel plug-in for predictions or you can use code. How to write DMX code for predictions: http://technet.microsoft.com/en-us/library/ms132031.aspx

Thanks for the feedback about the viewer.

If you have more/ additional questions, let me know.

Regards,
Joris

New Post: Tutorial?

pauloebc — Mon, 10 Nov 2008 12:07:10 GMT

Joris Volkonet,

I'm an statician, phD in Markovs Chain. Note that I'm creating a process in SQL Data Mining and the modeling, is the final part of the process flow, because I need to run an algorithm to predict good/bad. But neural networks as are not stable for that, which means that if I take a 2 samples from the same dataset, I'll get two different models (ranked by score tiers). Note that from the statistic point of view this is terrible.

So I think that SVM is a great solution for that issue.

But as I said, I dont know a lot how it works.

1. Could give a hand telling how to configure the algorithm to predict a variable good/bad from a list of inputs?

2. How many rows approximally supports the algorithm?

3. How many inputs supports (how many dimentions)?

4. Whats the fuction do I need to use to put the predict in a table?

ps: the viewer is working, but just on SQL Data Mining 2008

I'll be glad if you could help me.

Great job!

Best regards,

Paulo Carvalho

On Mon, Nov 10, 2008 at 7:58 AM, JorisValkonet <notifications@codeplex.com> wrote:

From: JorisValkonet

Hi,

we will work on a tutorial about the algorithm. This might take some time. Therefore, I will try to answer your questions here.

Kernels and variables

Variable Description

C With the C parameter you can control the trade-off between the margin and misclassification. The lower the C parameter, the less misclassifications should occur, but it also could lead to overfitting. Values [0, infinite] are floats.

kernel type You can specify the kernel the algorithm should use. Possible values: [linear, polynomial, rbf]. Default: [linear]

Cache size You can specify the size of the kernel cache used by the algorithm. The default value should be good and the type is int.

Gamma The gamma parameter is used by the RBF kernel. Values [0, infinite] are floats.

Exponent The exponent parameter is used by the polynomial kernel. It is the power parameter in the equation. The values should be floats.

Inhomogeneous The Inhomogeneous parameter is used by the polynomial kernel. The values are [true, false] and indicate which type of polynomial kernel should be used.

More details about the kernels can be found on many places on the internet, like wikipedia. We will describe the kernels in more detail in the tutorial.

The viewer
I cannot tell you why the viewer fails without more details. Can you tell me on whether you installed it on SQL2005 or SQL2008. Are you trying to use the viewer from Excel, BIDS or Management Studio? Do you get an error message from the viewer or installer? The problem could be in the registry, the location where the dll is placed or the compilation type of the dll.

Results
It depends on what you mean by the results. There are supported functions by the algorithm which provide additional information but these functions are not really results of the algorithm, but more descriptive about the metadata of the input. Unfortunately, there is no function implemented in this first version to return the weight vector calculated by the SMO algorithm. If you like the results of the predictions be put into a table, then there are greate resources on www.sqlserverdatamining.com.

Read the full discussion online.

To add a post to this discussion, reply to this email (svmplugin@discussions.codeplex.com)

To start a new discussion for this project, email svmplugin@discussions.codeplex.com

You are receiving this email because you subscribed to this discussion on CodePlex. You can unsubscribe on codePlex.com.

Please note: Images and attachments will be removed from emails. Any posts to this discussion will also be available online at codeplex.com

New Post: Tutorial?

JorisValkonet — Mon, 10 Nov 2008 09:57:51 GMT

Hi,

we will work on a tutorial about the algorithm. This might take some time. Therefore, I will try to answer your questions here.

Kernels and variables

Variable	Description
C	With the C parameter you can control the trade-off between the margin and misclassification. The lower the C parameter, the less misclassifications should occur, but it also could lead to overfitting. Values [0, infinite] are floats.
kernel type	You can specify the kernel the algorithm should use. Possible values: [linear, polynomial, rbf]. Default: [linear]
Cache size	You can specify the size of the kernel cache used by the algorithm. The default value should be good and the type is int.
Gamma	The gamma parameter is used by the RBF kernel. Values [0, infinite] are floats.
Exponent	The exponent parameter is used by the polynomial kernel. It is the power parameter in the equation. The values should be floats.
Inhomogeneous	The Inhomogeneous parameter is used by the polynomial kernel. The values are [true, false] and indicate which type of polynomial kernel should be used.

More details about the kernels can be found on many places on the internet, like wikipedia. We will describe the kernels in more detail in the tutorial.

The viewer
I cannot tell you why the viewer fails without more details. Can you tell me on whether you installed it on SQL2005 or SQL2008. Are you trying to use the viewer from Excel, BIDS or Management Studio? Do you get an error message from the viewer or installer? The problem could be in the registry, the location where the dll is placed or the compilation type of the dll.

Results
It depends on what you mean by the results. There are supported functions by the algorithm which provide additional information but these functions are not really results of the algorithm, but more descriptive about the metadata of the input. Unfortunately, there is no function implemented in this first version to return the weight vector calculated by the SMO algorithm. If you like the results of the predictions be put into a table, then there are greate resources on www.sqlserverdatamining.com.

New Post: Tutorial?

pauloebc — Fri, 07 Nov 2008 09:50:16 GMT

I'd like to know about the functionalaty of the algorithm, like:

How many variable the algorithm support?

What kind of variable (int, double, float)?

In my machine, the plugin works but the viewer fails, why?

How can I put the results into new table? whats the fuction do I need to use (GetAttribute...etc)?

New Post: Functionality for the next version

JorisValkonet — Sun, 26 Oct 2008 13:11:01 GMT

Please tell us what we should add in the next version of the plug-in...