Model file--

svpriyansvpriyan Member Posts: 29 Maven
edited November 2018 in Help
Dear Sir,
I am Priyan here,
My assignment is  basically Market Basket Analysis so I started with these steps

try to use FP growth Algorithm ; but here i got an error  '

A value for the parameter 'model_file' must be specified
So here what I can specify as the Model .. ??? .mod
How i can see this file '? or make this file for my exercise.. is it possible to make it
Thanz
Priyan

Answers

  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi Priyan,
    could you describe your problem a little bit more detailed,please? And perhabs post your process here, that will help us understand your problem.

    Greetings,
      Sebastian
  • svpriyansvpriyan Member Posts: 29 Maven
    Thanks for the reply; I was unable to reply on time
    Actually I am doing Market basket analysis as my assignment.
    I need to use FP growth Algorithm to do so.
    I followed these steps and able to find some Frequent item sets.
    + Root
            + Input
            + UserBasedDiscretization
            + FPGrowth
            +AssociationRuleGenerator
          + Result writer
    And I able to get the result with the data ‘golf-store.aml’
    But I have to do with my data, stored in the DB2 database. I don’t know how to start with. I need to  know your advice. On it..
    Greetings
    Priyan
  • TobiasMalbrechtTobiasMalbrecht Moderator, Employee, Member Posts: 294 RM Product Management
    Hi,

    ok, lets start at the beginning:

    1. Have you been able to conect to your DB2 database and read your data into RapidMiner?
    2. What does your data look like? How many examples? Attributes? What types of attributes?
    3. Are there any error messages?

    Regards,
    Tobias
  • svpriyansvpriyan Member Posts: 29 Maven
    Hai,
    I have problems in connecting the database even. I used DatabaseExampleSource to connect DB2.

    It is asking for Database Server/ IP. Though i gave the IP but it is asking for user name/ pwd.  Else where i can find the Database Server Name?
    I have the login user name & password of the DB, but it gives the message as '..............revoked....'


    So I need to know what can i do with these Database server Name, user name , password. 
    thanks


    [attachment deleted by admin]
  • IngoRMIngoRM Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder
    Hi,

    first thing: according to the screenshot you used the operator DatabaseExampleSetWriter instead of the operator DatabaseExampleSetSource. Next question: did you install the necessary driver for DB2 (delivered with the system)?

    Cheers,
    Ingo
  • svpriyansvpriyan Member Posts: 29 Maven
    Hai,
    I used what you said..
    i have attached the Screen shot about ODBC connection drivers.  tell me what i can choose..
    secondly, I need to know what is the name i can give when it as for the Server name & IP,  So what about User Name & password.

    thanks




    [attachment deleted by admin]
  • TobiasMalbrechtTobiasMalbrecht Moderator, Employee, Member Posts: 294 RM Product Management
    Hi,
    svpriyan wrote:

    i have attached the Screen shot about ODBC connection drivers.  tell me what i can choose..
    secondly, I need to know what is the name i can give when it as for the Server name & IP,  So what about User Name & password.
    in the screen you attached you can only see what database drivers are available. You can not select anything, it is just for your information. In the following dialog step you are asked for your database driver, the host name / IP and the database name.

    For the hose name / IP you have to (as it implies) specify the name of the host or the IP of your database server. The database name is the name of your database in your database system. As user and password you have to set the user information that you need to access your database.

    Regards,
    Tobias
  • svpriyansvpriyan Member Posts: 29 Maven
    thanks I got it.. though i have to ask the following errors.
    Error 1: The table name should be what  "work_on_database". should be different??


    Error 2:

    P Nov 25, 2008 3:29:09 PM: Checking properties...
    P Nov 25, 2008 3:29:09 PM: Properties are ok.
    P Nov 25, 2008 3:29:09 PM: Checking process setup...
    P Nov 25, 2008 3:29:09 PM: Inner operators are ok.
    P Nov 25, 2008 3:29:09 PM: Checking i/o classes...
    P Nov 25, 2008 3:29:09 PM: [Error] DatabaseExampleSetWriter: DatabaseExampleSetWriter: DatabaseExampleSetWriter: Missing input: ExampleSet
    P Nov 25, 2008 3:29:09 PM: [Error] There was 1 error.

    I dont have any place to give the path to my table... thats why this error shows up.. is it.. where i can give my table.., I have 3-5 tables. in my database.

    [attachment deleted by admin]
  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    Ingo already pointed out the solution for your errors. See http://rapid-i.com/rapidforum/index.php/topic,407.msg1667.html#msg1667

    Greetings,
      Sebastian
  • svpriyansvpriyan Member Posts: 29 Maven
    Hello,
    Thanks for the previous replies.

    This is my data format

    Price DECIMAL 0- infinite
    Amount DECIMAL 1-500
    Name CHAR  6
    Desc CHAR 10
    Values DECIMAL Any values between 0-1000
    Ex Values DECIMAL Any values between 0-1000


    I wanted to do preprocessing so I used UserBasedDiscretization and I found
    Difficulties how to define classes and Upper limits for each of these( except Name & desc- we don’t need it)
    Is that any way to do this or do we need to use any other way to preprocess it.
    i have 150000 records for this relation.

    Thanks



  • TobiasMalbrechtTobiasMalbrecht Moderator, Employee, Member Posts: 294 RM Product Management
    Hi,
    svpriyan wrote:

    I wanted to do preprocessing so I used UserBasedDiscretization and I found
    Difficulties how to define classes and Upper limits for each of these( except Name & desc- we don’t need it)
    Is that any way to do this or do we need to use any other way to preprocess it.
    how to specify the intervals in the [tt]UserBasedDiscretization[/tt] or whether to use it or any other discretization of course strongly depends on your data but also on the task you wish to accomplish. It is also extremely relevant what you intend to do with the data after preprocessing or discretization in particular. Hence, you should be a little bit more specific on your analysis goal ...

    Regards,
    Tobias
  • svpriyansvpriyan Member Posts: 29 Maven
    Thanks for the Questions.

    Basically I am going to do the market basket Analysis, so initially i am going to do the discretization and then plan to use FPGrowth algorithm and then Association Rule Generation. to get identify some patterns.
    Can i get some idea then from your side.
    Thanks
    Priyan
  • TobiasMalbrechtTobiasMalbrecht Moderator, Employee, Member Posts: 294 RM Product Management
    Hi Priyan,

    ok, (hopefully) last question: you have transaction data in the form you need for FPGrowth? I could not quite get that from the data description you posted. It seems to me, you have data where each example just considers one article, am I right? Then it will not be sufficient to use [tt]UserBasedDiscretization[/tt]

    For FPGrotwth you need transactional data like that, containing binary attributes for each article indicating whether an article was part in the transaction:

    transaction_id      article1      article2
    1                            1                0
    2                            1                1
    3                            1                0
    4                            0                1

    Regards,
    Tobias
  • svpriyansvpriyan Member Posts: 29 Maven
    hai,
    yes i have transactional data ..( see the image)
    so how can i start all these ?
    so how can i do discretization ? how can i solve this.. lets start from here .

    Thanks

    priyan

    [attachment deleted by admin]
  • TobiasMalbrechtTobiasMalbrecht Moderator, Employee, Member Posts: 294 RM Product Management
    Hi Priyan,

    well from the data snippet you posted I can not fully understand the structure. Even the names from the previous post does not help me here. So could you please describe (in detail!) what the columns and the examples are in the data.

    The reason why I am asking is that I still do not get how your information is stored. As you probably know, fpgrowth will search for frequent item sets. But that does imply, you have transactional data containing the information which articles belong to a transaction at the same time. Hence, it does not help, that cola was bought in 126 transactions and chips were bought in 254 transactions. Instead the data must contain in which transactions both were bought together. Unfortunately I can not conclude on how to get such a format from the data snippet you posted ...

    Regards,
    Tobias
Sign In or Register to comment.