4 Installing and Using the Sample Programs

The PL/SQL Samples

The files with a sql extension are sample programs that illustrate the PL/SQL API and SQL scoring functions, administrative scripts, and files related to the BLAST algorithm.

Table 4-1 lists the PL/SQL programs that illustrate the mining functions and algorithms supported by Oracle Data Mining.

Table 4-2 lists the administrative scripts related to the data mining sample programs.

Table 4-3 lists the files related to the BLAST algorithm. Instructions for creating sample data sets and running the BLAST sample program are provided in "Using the BLAST Sample".

The Java Samples

The files with a java extension are sample programs that illustrate the Java API. In addition to the programs that illustrate the basic mining functions, there are programs that illustrate predictive analytics and specific mining tasks, such as data transformations and model import/export.

Table 4-4 lists the Java programs that illustrate the mining functions and algorithms supported by Oracle Data Mining.

Table 4-5 lists the Java programs that illustrate special mining tasks. These features are all supported in the PL/SQL API as well, since the Java API is layered on the PL/SQL API.

A Sample User Scenario

Let's assume that you have access to Oracle Database 10g, which has been installed on a Windows host with the Data Mining option, the sample schemas, and the sample programs. You can log in to this database as SYS using SQL*Plus.

Several Data Mining users (dmuser1 and dmuser2) have already been created in this database. You want to create your own user ID (dmuser3) and run the sample programs to create the sample models in your schema. To accomplish this, you would perform the following steps.

1. Log in to the database as SYS and create the dmuser3 user.

   > sqlplus
       Enter user-name: sys / as sysdba
       Enter password: sys_password
   SQL> create user dmuser3 identified by dmuser3_password 
        default tablespace users temporary tablespace temp 
        quota unlimited on users;
   
   

2. Run dmshgrants.sql to grant privileges to dmuser3. These privileges are required for performing data mining tasks and for accessing the mining data in SH used by the sample programs.

   SQL> @ %ORACLE_HOME%\rdbms\demo\dmshgrants SH_password dmuser3
   
   

3. Connect as dmuser3 and run dmsh.sql to populate the dmuser3 schema with views and tables needed by the sample programs. Save your changes in the database.

   SQL> connect dmuser3/dmuser3_password
   SQL> @ %ORACLE_HOME%\rdbms\demo\dmsh
   SQL> commit;
   
   

You can now run any of the PL/SQL data mining samples as described in "Using the PL/SQL Samples". For example, while logged in to SQL*Plus as dmuser3, you could run the Naive Bayes PL/SQL sample with the following statement.

SQL>@ %ORACLE_HOME%\rdbms\demo\dmnbdemo

You can also run any of the Java data mining samples, if you have Java 1.4.2 or higher and your CLASSPATH is set as described in "Using the Java Samples". For example, you could run the Naive Bayes Java sample with a command like the following at the operating system prompt.

>java dmnbdemo myserver:1521:orcl dmuser3 dmuser3_password

Note that the BLAST demo uses different data sets and requires a separate setup procedure. See "Using the BLAST Sample".

Customer Data for Data Mining

Views in the data mining user's schema define columns of data from the CUSTOMERS, SALES, PRODUCTS, COUNTRIES, and SUPPLEMENTARY_DEMOGRAPHICS tables in the SH schema. You can list these views with the following SQL statements.

SQL>connect dmuser3/dmuser3_password
SQL>select view_name from user_views;

The views are listed in Table 4-6.

You can see the references to tables in SH by listing the view definitions. The definition of the view MINING_DATA_BUILD_V is shown as follows.

SQL>set long 1000000
SQL>set longc 100000
SQL>select text from all_views where 
    owner='DMUSER3'and view_name='MINING_DATA_BUILD_V';

      SELECT a.CUST_ID, a.CUST_GENDER, 2003-a.CUST_YEAR_OF_BIRTH AGE, 
             a.CUST_MARITAL_STATUS, c.COUNTRY_NAME, a.CUST_INCOME_LEVEL,
             b.EDUCATION, b.OCC UPATION, b.HOUSEHOLD_SIZE, b.YRS_RESIDENCE,
             b.AFFINITY_CARD, b.BULK_PACK_DISKETTES, b.FLAT_PANEL_MONITOR,
             b.HOME_THEATER_PACKAGE, b.BOOKKEEPING _APPLICATION, 
             b.PRINTER_SUPPLIES, b.Y_BOX_GAMES, b.OS_DOC_SET_KANJI 
       FROM  sh.customers a, 
             sh.supplementary_demographics b, 
             sh.countries c 
       WHERE a.CUST_ID = b.CUST_ID AND a.country_id = c.country_id 
             AND a.cust_id between 101501 and 103000 

The views are used to build, test, and score the sample models. Each view has a CUSTOMER_ID column, which is the case ID, and an AFFINITY_CARD column, which is the target used by the predictive models. Most of the views provide data for 1500 customers (1500 rows). The view used by the one-class SVM model has data for 940 customers.

The columns of build data in the MINING_DATA_BUILD_V view are listed in the following example.

SQL> describe MINING_DATA_BUILD_V

CUST_ID                    NOT NULL            NUMBER
CUST_GENDER                NOT NULL            CHAR(1)
AGE                                            NUMBER
CUST_MARITAL_STATUS                            VARCHAR2(20)
COUNTRY_NAME               NOT NULL            VARCHAR2(40)
CUST_INCOME_LEVEL                              VARCHAR2(30)
EDUCATION                                      VARCHAR2(21)
OCCUPATION                                     VARCHAR2(21)
HOUSEHOLD_SIZE                                 VARCHAR2(21)
YRS_RESIDENCE                                  NUMBER
AFFINITY_CARD                                  NUMBER(10)
BULK_PACK_DISKETTES                            NUMBER(10)
FLAT_PANEL_MONITOR                             NUMBER(10)
HOME_THEATER_PACKAGE                           NUMBER(10)
BOOKKEEPING_APPLICATION                        NUMBER(10)
PRINTER_SUPPLIES                               NUMBER(10)
Y_BOX_GAMES                                    NUMBER(10)
OS_DOC_SET_KANJI                               NUMBER(10)

Market Basket Data for Association Rules

The association rules samples use the MARKET_BASKET_V data set, which includes columns of products from the PRODUCTS table and the CUSTOMER_ID column from the CUSTOMERS table in SH. The columns of the MARKET_BASKET_V view are listed in the following example.

SQL> describe MARKET_BASKET_V

CUST_ID                     NOT NULL         NUMBER
EXTENSION_CABLE                              NUMBER
FLAT_PANEL_MONITOR                           NUMBER
CD_RW_HIGH_SPEED_5_PACK                      NUMBER
ENVOY_256MB_40GB                             NUMBER
ENVOY_AMBASSADOR                             NUMBER
EXTERNAL_8X_CD_ROM                           NUMBER
KEYBOARD_WRIST_REST                          NUMBER
SM26273_BLACK_INK_CARTRIDGE                  NUMBER
MOUSE_PAD                                    NUMBER
MULTIMEDIA_SPEAKERS_3INCH                    NUMBER
OS_DOC_SET_ENGLISH                           NUMBER
SIMM_16MB_PCMCIAII_CARD                      NUMBER
STANDARD_MOUSE                               NUMBER 

Customer Data for Text Mining

The text mining samples use the same customer data from tables in SH, but they include either an extra text column or a collection type column. The collection type is a nested table of type DM_NESTED_NUMERICALS.

You can list these tables with the following SQL statements.

SQL>connect dmuser3/dmuser3_password
SQL>select table_name from user_tables where table_name like '%MINING%';

The tables are listed in Table 4-7.

In the MINING_BUILD_TEXT, MINING_TEST_TEXT, and MINING_APPLY_TEXT tables, the COMMENTS column is of type VARCHAR2(4000).

SQL> describe MINING_BUILD_TEXT
 Name                                      Null?    Type
 ----------------------------------------- -------- ----------------------------
 CUST_ID                                   NOT NULL NUMBER
 CUST_GENDER                               NOT NULL CHAR(1)
 AGE                                                NUMBER
 CUST_MARITAL_STATUS                                VARCHAR2(20)
 COUNTRY_NAME                              NOT NULL VARCHAR2(40)
 CUST_INCOME_LEVEL                                  VARCHAR2(30)
 EDUCATION                                          VARCHAR2(21)
 OCCUPATION                                         VARCHAR2(21)
 HOUSEHOLD_SIZE                                     VARCHAR2(21)
 YRS_RESIDENCE                                      NUMBER
 AFFINITY_CARD                                      NUMBER(10)
 BULK_PACK_DISKETTES                                NUMBER(10)
 FLAT_PANEL_MONITOR                                 NUMBER(10)
 HOME_THEATER_PACKAGE                               NUMBER(10)
 BOOKKEEPING_APPLICATION                            NUMBER(10)
 PRINTER_SUPPLIES                                   NUMBER(10)
 Y_BOX_GAMES                                        NUMBER(10)
 OS_DOC_SET_KANJI                                   NUMBER(10)
 COMMENTS                                           VARCHAR2(4000)

In the MINING_*_NESTED_TEXT tables, the COMMENTS column is of type DM_NESTED_NUMERICALS.

SQL> describe MINING_BUILD_NESTED_TEXT
 Name                                      Null?    Type
 ----------------------------------------- -------- ----------------------------
 CUST_ID                                   NOT NULL NUMBER
 CUST_GENDER                               NOT NULL CHAR(1)
 AGE                                                NUMBER
 CUST_MARITAL_STATUS                                VARCHAR2(20)
 COUNTRY_NAME                              NOT NULL VARCHAR2(40)
 CUST_INCOME_LEVEL                                  VARCHAR2(30)
 EDUCATION                                          VARCHAR2(21)
 OCCUPATION                                         VARCHAR2(21)
 HOUSEHOLD_SIZE                                     VARCHAR2(21)
 YRS_RESIDENCE                                      NUMBER
 AFFINITY_CARD                                      NUMBER(10)
 BULK_PACK_DISKETTES                                NUMBER(10)
 FLAT_PANEL_MONITOR                                 NUMBER(10)
 HOME_THEATER_PACKAGE                               NUMBER(10)
 BOOKKEEPING_APPLICATION                            NUMBER(10)
 PRINTER_SUPPLIES                                   NUMBER(10)
 Y_BOX_GAMES                                        NUMBER(10)
 OS_DOC_SET_KANJI                                   NUMBER(10)
 COMMENTS                                           DMSYS.DM_NESTED_NUMERICALS

The process of extracting terms from a text column into a nested table column is described in "Using the Text Mining Samples" and in the Oracle Data Mining Application Developer's Guide.

Running the PL/SQL Programs

In SQL*Plus, use commands like the following to execute the sample programs and list the models created by them.

>sqlplus dmuser3/dmuser3_password
SQL> set serveroutput on
SQL> set echo on
SQL> @ %ORACLE_HOME%\rdbms\demo\program_name
.
.
.
.
SQL> set linesize 200
SQL> set pagesize 100 
SQL> select NAME, FUNCTION_NAME, ALGORITHM_NAME, TARGET_ATTRIBUTE from DM_USER_MODELS; 
 
NAME                   FUNCTION_NAME           ALGORITHM_NAME                  TARGET_ATTRIBUTE
---------------------  ----------------------  ------------------------------  -----------------
T_NMF_SAMPLE           FEATURE_EXTRACTION      NONNEGATIVE_MATRIX_FACTOR                    
T_SVM_CLAS_SAMPLE      CLASSIFICATION          SUPPORT_VECTOR_MACHINES         AFFINITY_CARD
AR_SH_SAMPLE           ASSOCIATION_RULES       APRIORI_ASSOCIATION_RULES                    
AI_SH_SAMPLE           ATTRIBUTE_IMPORTANCE    MINIMUM_DESCRIPTION_LENGTH      AFFINITY_CARD
ABN_SH_CLAS_SAMPLE     CLASSIFICATION          ADAPTIVE_BAYES_NETWORK          AFFINITY_CARD
DT_SH_CLAS_SAMPLE      CLASSIFICATION          DECISION_TREE                   AFFINITY_CARD
NB_SH_CLAS_SAMPLE      CLASSIFICATION          NAIVE_BAYES                     AFFINITY_CARD
SVMC_SH_CLAS_SAMPLE    CLASSIFICATION          SUPPORT_VECTOR_MACHINES         AFFINITY_CARD
OC_SH_CLUS_SAMPLE      CLUSTERING              O_CLUSTER                                    
KM_SH_CLUS_SAMPLE      CLUSTERING              KMEANS                                       
NMF_SH_SAMPLE          FEATURE_EXTRACTION      NONNEGATIVE_MATRIX_FACTOR                    
SVMR_SH_REGR_SAMPLE    REGRESSION              SUPPORT_VECTOR_MACHINES         AGE          

PL/SQL Program Summaries

Summary descriptions of the PL/SQL sample programs are provided in Table 4-8. The programs are listed individual by name and function in Table 4-1.

For detailed descriptions of the sample programs, see the comments in the source code.

SQL Scoring Functions

Some of the PL/SQL sample programs use SQL scoring functions to apply models created with the DBMS_DATA_MINING package. The SQL scoring functions can also be used to apply models created with the Java API.

The programs that demonstrate the SQL scoring functions are listed in Table 4-9.

Preparing to Run the Java Programs

Before running the Java programs, do the following:

1. Check that the version of Java you are using is 1.4.2. You can execute the following in a command window to check the version of Java:

   >java -version
   
   

2. Ensure that your CLASSPATH includes the following JAR files:

   %ORACLE_HOME%\rdbms\jlib\jdm.jar
               %ORACLE_HOME%\rdbms\jlib\ojdm_api.jar
               %ORACLE_HOME%\rdbms\jlib\xdb.jar
               %ORACLE_HOME%\jdbc\lib\ojdbc14.jar
               %ORACLE_HOME%\oc4j\j2ee\home\lib\connector.jar
               %ORACLE_HOME%\jlib\orai18n.jar   
               %ORACLE_HOME%\jlib\orai18n-mapping.jar
               %ORACLE_HOME%\lib\xmlparserv2.jar
   
   

3. Compile the programs. Use the javac executable in the \jdk\bin\ directory within Oracle home.

   >javac program_name
   

Running the Java Programs

Use the following syntax to execute the sample Java programs. Use the java executable in the \jdk\bin\ directory within Oracle home.

>java program_name connect_string user_name user_password

The connection string specifies your Oracle database connection. It identifies the machine hosting the database, the port through which the connection is made, and the name of the database instance (the Oracle system identifier).

host_name:port:SID

For example, the following command executes the Naive Bayes sample dmnbdemo.java as dmuser3 in the database instance orcl on host mach05 at port 1521.

>java dmnbdemo mach05:1521:orcl dmuser3 dmuser3_password

You can list the models created by the Java programs with commands like the following in SQL*Plus.

>sqlplus dmuser3/dmuser3_password
SQL> set linesize 200
SQL> set pagesize 100 
SQL> select NAME, FUNCTION_NAME, ALGORITHM_NAME, TARGET_ATTRIBUTE 
            from dm_user_models where NAME like '%JDM'; 

NAME                FUNCTION_NAME          ALGORITHM_NAME                 TARGET_ATTRIBUTE
-----------------   ---------------------  ----------------------------   ------------------
TXTNMFMODEL_JDM     FEATURE_EXTRACTION     NONNEGATIVE_MATRIX_FACTOR
ARMODEL_JDM         ASSOCIATION_RULES      APRIORI_ASSOCIATION_RULES
TREEMODEL_JDM       CLASSIFICATION         DECISION_TREE                  AFFINITY_CARD
AIMODEL_JDM         ATTRIBUTE_IMPORTANCE   MINIMUM_DESCRIPTION_LENGTH     AFFINITY_CARD
ABNMODEL_JDM        CLASSIFICATION         ADAPTIVE_BAYES_NETWORK         AFFINITY_CARD
NBMODEL_JDM         CLASSIFICATION         NAIVE_BAYES                    AFFINITY_CARD
SVMCMODEL_JDM       CLASSIFICATION         SUPPORT_VECTOR_MACHINES        AFFINITY_CARD
SVMOMODEL_JDM       CLASSIFICATION         SUPPORT_VECTOR_MACHINES
KMMODEL_JDM         CLUSTERING             KMEANS
OCMODEL_JDM         CLUSTERING             O_CLUSTER
NMFMODEL_JDM        FEATURE_EXTRACTION     NONNEGATIVE_MATRIX_FACTOR
SVMRMODEL_JDM       REGRESSION             SUPPORT_VECTOR_MACHINES        AGE
TXTSVMMODEL_JDM     CLASSIFICATION         SUPPORT_VECTOR_MACHINES        AFFINITY_CARD

Java Program Summaries

Summary descriptions of the Java sample programs are provided in Table 4-10. The programs are listed individually by name in Table 4-4 and Table 4-5.

For detailed descriptions of the sample programs, see the comments in the source code.

Text Mining in PL/SQL

Three PL/SQL sample programs illustrate the process of text mining. One program illustrates the pre-processing that is required to prepare the data for mining. The other two programs build models that use the transformed text.

MINING_BUILD_NESTED_TEXT
MINING_TEST_NESTED_TEXT
MINING_APPLY_NESTED_TEXT

SQL> @ %ORACLE_HOME%\rdbms\demo\dmtxtnmf.sql
SQL> @ %ORACLE_HOME%\rdbms\demo\dmtxtsvm.sql
SQL> select NAME, FUNCTION_NAME, ALGORITHM_NAME, TARGET_ATTRIBUTE  
           from dm_user_models;NAME               FUNCTION_NAME      ALGORITHM_NAME            TARGET_ATTRIBUTE ----------------   ------------------ ------------------------  ----------------T_NMF_SAMPLE       FEATURE_EXTRACTION NONNEGATIVE_MATRIX_FACTOR T_SVM_CLAS_SAMPLE  CLASSIFICATION     SUPPORT_VECTOR_MACHINES   AFFINITY_CARD

Text Mining in Java

Two Java sample programs illustrate the process of text mining. One builds a feature extraction model, the other builds a classification model.

MINING_BUILD_TEXT
MINING_TEST_TEXT
MINING_APPLY_TEXT

> java dmtxtnmfdemo host:port:SID dmuser3 dmuser3_password
> java dmtxtsvmdemo host:port:SID dmuser3 dmuser3_password
>sqlplus dmuser3/dmuser3_password
SQL> select NAME, FUNCTION_NAME, ALGORITHM_NAME, TARGET_ATTRIBUTE 
           from dm_user_models;NAME               FUNCTION_NAME       ALGORITHM_NAME            TARGET_ATTRIBUTE ----------------   ------------------  ------------------------  ----------------txtnmfModel_jdm    FEATURE_EXTRACTION  NONNEGATIVE_MATRIX_FACTOR txtsvmModel_jdm    CLASSIFICATION      SUPPORT_VECTOR_MACHINES   AFFINITY_CARD

Preparing to Run the BLAST Demo

The sample BLAST table functions in dmbldemo.sql use two data sets: SwissProt and ecoli10. To prepare these data sets, log in to SQL*Plus as the data mining user and run the dmblprot and dmblcoli scripts as shown in the following example.

SQL>connect dmuser3/dmuser3_password
SQL>@ %ORACLE_HOME%\rdbms\demo\dmblprot.sql
SQL>@ %ORACLE_HOME%\rdbms\demo\dmblcoli.sql

Exit SQL*Plus and use the SQL*Loader utility to load data into the SwissProt database in the schema of the data mining user. From the command prompt, change to the \rdbms\demo directory under Oracle home and execute the following command.

>sqlldir dmuser3/dmuser3_password 
         control=dmblprot.ctl data=dmblprot.txt log=dmblprot.log

Running the BLAST Table Functions

The sample program dmbldemo.sql contains multiple invocations of BLAST table functions. You can run them all at once by running the dmbldemo.sql script, or you can copy individual table functions to the SQL*Plus command line and execute them individually.

To run the sample program, log in to SQL*Plus as the data mining user and run the dmbldemo.sql script as shown in the following example.

SQL>connect dmuser3/dmuser3_password
SQL> @ %ORACLE_HOME%\rdbms\demo\dmbldemo.sql

The following example shows how you could execute the sample BLASTP_MATCH table function at the SQL*Plus command line. BLASTP_MATCH is run against all the human proteins in the SwissProt database.

SQL>connect dmuser3/dmuser3_password
SQL>SET serveroutput ON
SQL>SET trimspool ON  
SQL>SET pages 10000
SQL>column seq_id format a10
SQL>select T_SEQ_ID AS seq_id, score, EXPECT as evalue
         from TABLE(
              BLASTP_MATCH (
                   (select sequence from query_db),
                    CURSOR(SELECT seq_id, seq_data 
                           FROM swissprot
                           WHERE organism = 'Homo sapiens (Human)'),
                   1,
                  -1,
                   0,
                   0,
                  'BLOSUM62',
                  10,
                   0,
                   0,
                   0,
                   0,
                   0)
          )
     order by score,seq_id;

The output of this table function is shown as follows.

SEQ_ID          SCORE     EVALUE
---------- ---------- ----------
P58107             49 7.24297332
P31947            169 8.8130E-14
P27348            198 3.8228E-17
Q04917            198 3.8228E-17
P31946            205 5.8977E-18