data mining lift chart problem on vs2005

Hello . Because of my graduation project , I interested in data mining application , Adventureworks DW on MS VS 2005 . I opened File->Open->project/solution ->Enterprise -> AdventureworksDW .then I successfully deployed the algorithms decision tree and Clustering . Then I opened tab Mining Accuracy Chart then selected input table "testing" , which I had created before , from vTargetMail . After that , mining structure table and target mail table has automaticaly linked each other .Next , I selected predictive input as 1 , of the predictable row "BikeBuyer" . But , when I clicked "Lift Chart ", I only got a 45 degree line , everytime .. How can I fix it ?

It depends - how many rows are in your test set? What do the threes and clusters look like - do they make sense? Maybe there's something wrong with your test data? What shows up in the chart legend - the models you created? What shows up in the floating/docked mining legend when you click on the chart?

I'll need some more information to help you with the issue.

Thanks

-Jamie

|||

I have table LG_131_01 having 2500 rows and I am using algorithms as shown below .

Clustering Algorithm:

Decision Tree Algorithm:

Naive Bayes Algorithm:

and Association Rules Algorithm:

But when I click to "Lift Chart"

I got the following error ..

Failed to execute the query due to the following error:

XML for Analysis parser: The XML for Analysis request timed out before it was completed.
Execution of the managed stored procedure GenerateLiftTableUsingDatasource failed with the following error:

Exception has been thrown by the target of an

invocation.Microsoft::AnalysisServices::AdomdServer::AdomdException.

Please help ..

|||I am still waiting , please help ..|||

The timeout exception issues can be fixed by changing the query execution timeout.

In BI Dev Studio, the Tools menu, select Options, then Business Intelligence Designers. There, change the Query Timeout property. The value should be specified in seconds. A value of 3600 (one hour) should be enough even for the most complex models

Hope this helps

|||

Sir I did your suggestion , it suceed . But now I only got the 45 degree line , comparing Decision trees alg and clustering algorithms . Is it strange ? Why do I get ideal model line like as two algs ? As seen above , testing results were fine .. And I have got table having enough rows of data 2591 rows of data ... Or does the program has a bug ?

|||

The only time I've seen this happen is with using a bad testing set (e.g. using the predicted results as a testing set).

What happens when you use the training set as the testing set? Or try the other form of the lift chart (using a specific state)?

Data manipulation with One-to-many-to-one relationship

Is it possible to INSERT, UPDATE, DELETE data in this type of relationship? If so, how in VS2005? Not having issues with SELECT, even without joins.

Thanks

Hi,

Try providing more information about what you are trying to do. Do you have tables set up in a DB already? What columns do you have in those tables, and what are the references that make a given row of data in each table unique (primary keys)?

If you don't have a schema yet, provide (in your own words), what facts you are trying to convey - such as 'A customer places an order. Order has line items. Each product has a product number.' etc... This will give viewers here a better understanding of what you need.

Good luck, in any event. BRN..

|||

Let's say for example:

Table 1 has Table1ID (PK) and Table1Data

Table 2 has Table2ID (PK), Table1ID (FK for Table1.Table1ID), Table3ID (FK for Table3.Table3ID)

Table 3 has Table3ID (PK), and Table3Data

Relationship is one-to-many for Table1.Table1ID to Table2.Table1ID and one-to-many for Table3.Tat.ble3ID to Table2.Table3ID, creating a many to many relationship for Table1 to Table 3 using Table 2 as an intersection table. I do not have any cascading.

I am able to query the database and gather the information, but need to be able to add, edit, and delete in a C# VS2005 project. Starting off small with something like this would enable me to conquer a bigger project that I would like to create. This is the first time I have had to work with data that required this type of relationship.

I'd prefer to see the SQL.

Is this enough information?

Many thanks

|||OK, let take these 3 tables for example:

create table Employee (EmplNumber int not null primary key,EmplName nvarchar(20) not null)

create table Department (DeptNumber int not null primary key,DeptName nvarchar(50) not null)

-- relationship between Employee and Department

create table DeptEmpl (DeptNumber int not null foreign key references Department(DeptNumber)
ON DELETE CASCADE ON UPDATE CASCADE,
EmplNumber int not null foreign key references Employee(EmplNumber)
ON DELETE CASCADE ON UPDATE CASCADE)

insert into Employee values (1, 'John')
insert into Employee values (2, 'Terry')
insert into Employee values (3, 'Andrew')

insert into Department values (1, 'Marketing')
insert into Department values (2, 'Sales')
insert into Department values (3, 'Accounting')

insert into DeptEmpl values (1, 1)
insert into DeptEmpl values (1, 3)
insert into DeptEmpl values (2, 1)
insert into DeptEmpl values (2, 2)
insert into DeptEmpl values (2, 3)
insert into DeptEmpl values (3, 1)

?
There should be no problem when you try to insert data-- the PK/FK contraint will maintain the reference integrity for you. The key point here is theON DELETE/UPDATE CASCADE, they will help to maintain reference integrity during DELETE/UPDATE, so no extra work need to be done manually. If you have other concerns, feel free to post them|||

Thanks Iori_Jay

It is completely clear to me what you are doing in this example. How does one handle user input variables from a web application?

Let's say the user wants to insert a new Employee 'Edgar' that belongs to Department 'Sales'.

|||OK, typical steps for such INSERT should be:

1. Insert the information for the new employee 'Edgar' toEmployee table, and get theEmplNumber for the new inserted employee
2. Check theDepartment table to see whether there is a 'Sales' department, if not then add the 'Sales' department.
3. Insert a row toDeptEmpl table to represent the relationship between 'Edgar' and 'Sales'.

So a sample stored procedure looks like:

CREATE PROCEDURE usp_InsertNewEmployee @.EmpName nvarchar(20), @.DeptName nvarchar(50)
AS
IF((LEN(@.EmpName)=0) OR (LEN(@.DeptName)=0))
RAISERROR('Please input both Employee name and the department name which he(she) belongs to',16,1)
DECLARE @.newEmpNumber INT,@.DeptNumber INT
SELECT @.newEmpNumber=max(EmplNumber)+1 FROM Employee
SELECT @.DeptNumber=DeptNumber FROM Department WHERE DeptName=@.DeptName

INSERT INTO Employee SELECT @.newEmpNumber, @.EmpName
IF (@.DeptNumber IS NULL)
BEGIN
SELECT @.DeptNumber=max(DeptNumber)+1 FROM Department
INSERT INTO Department SELECT @.DeptNumber,@.DeptName
END
INSERT INTO DeptEmpl SELECT @.newEmpNumber,@.DeptNumber
go

EXEC usp_InsertNewEmployee 'Edgar','Sales'|||

Iori_jay,

Thanks so much for your help. I've learned many things from this post. I appreciate you sharing your talents.

ICE|8

|||It's my pleasure to help with this issue Wish you a happy festival!