Wednesday, January 4, 2017

Explicit Semantic Analysis in Oracle 12.2c Database

A new Oracle Data Mining algorithm in the Oracle 12.2c Database is called Explicit Semantic Analysis.

[The following examples are built using Oracle Data Miner 4.2 (SQL Developer 4.2) and the Oracle 12.2 Database cloud service (extreme edition) ]

The Explicit Semantic Analysis algorithm is an unsupervised algorithm used for feature extraction. ESA does not discover latent features but instead uses explicit features based on an existing knowledge base. There is no setup or install necessary to use this algorithm All you need is a licence for the Advanced Analytics Option for the database. The out from the algorithm is a distance measure that indicates how similar or dis-similar the input texts are, using the ESA model (and the training data set used). Let us look at an example. Setup training data for ESA Algorithm

Oracle Data Miner 4.2 (that comes with SQL Developer 4.2) has a data Wiki data set from 2005. This contains over 200,000 features. To locate the file go to.

...\sqldeveloper\dataminer\scripts\instWikiSampleData.sql

This file contains the DDL and the insert statements for the Wiki data set.

NewImage

After you run this script a new table called WIKISAMPLE table exists and contains records

NewImage

This gives us the base/seed data set to feed into the ESA algorithm.

Create the ESA Model using ODMr

To create the ESA model we have 2 ways of doing this. In this blog post I'll show you the easiest way by using the Oracle Data Miner (ODMr) tool. I'll have another blog post that will show you the SQL needed to create the model.

In an ODMr workflow create a new Data Source node. Then set this node to have the WIKISAMPLE table as it's data source.

Next you need to create the ESA node on the workflow. This node can be found in the Models section, of the Workflow Editor. The node is called Explicit Feature Extraction. Click on this node, in the model section, and then move your mouse to your workflow and click again. The ESA node will be created.

Join the Data Node to the ESA node by right clicking on the data node and then clicking on the ESA node.

Double click on the ESA node to edit the properties of the node and the algorithm.

NewImage

Explore the ESA Model and ESA Model Features

After the model node has finished you can now explore the results generated by the ESA model. Right click on the model node and select 'View Model'. The model properties window opens and it has 2 main tabs. The first of these is the coefficients tab. Here you can select a particular topic (click on the search icon beside the Feature ID) and select it from the list. The attributes and their coefficient values will be displayed.

NewImage

Next you can examine the second tab that is labeled as Features. In this table we can select a particular record and have a tag cloud and coefficients displayed. The tag cloud is a great way to see visually what words are important.

NewImage

How to use the ESA model to Compare new data using SQL

Now that we have the ESA model created, we can not use it model to compare other similar sets of documents.

You will need to use the FEATURE_COMPARE SQL function to evaluate the input texts, using the ESA model to compare for similarity. For example,

SELECT FEATURE_COMPARE(feat_esa_1_1
          USING 'Oracle Database is the best available for managing your data' text 
          AND USING 'The SQL language is the one language that all databases have in common' text) similarity 
FROM DUAL;
NewImage

The result we get is 0.7629.

The result generate by the query is a distance measure. The FEATURE_COMPARE function returns a comparison number in the range 0 to 1. Where 0 indicates that the text are not similar or related. If a 1 is returned then that indicated that the text are very similar or very related.

You can use this returned value to make a decision on what happens next. For example, it can be used to decide what the next step should be in your workflow and you can easily write application logic to manage this.

The examples given here are for general text. In the real world you would probably need a bigger data set. But if you were to use this approach in other domains, such as legal, banking, insurance, etc. then you would need to create a training data set based on the typical language that is used in each of those domains. This will then allow you to compare documents with each domain with greater accuracy.

[The above examples are built using Oracle Data Miner 4.2 (SQL Developer 4.2) and the Oracle 12.2 Database cloud service (extreme edition) ]

Wednesday, December 28, 2016

2016: A review of the year

As 2016 draws to a close I like to look back at what I have achieved over the year. Most of the following achievements are based on my work with the Oracle User Group community. I have some other achievements are are related to the day jobs (Yes I have multiple day jobs), but I won't go into those here.

As you can see from the following 2016 was another busy year. There was lots of writing, which I really enjoy and I'll be continuing with in 2017. As they say, watch this space for writing news in 2017.

Books

Yes 2016 was a busy year for writing and most of the later half of 2015 and the first half of 2016 was taken up writing two books. Yes two books. One of the books was on Oracle R Enterprise and this book compliments my previous published book on Oracle Data Mining. I now have the books that cover both components of the Oracle Advanced Analytics Option.

I also co-wrote a book with legends of Oracle community. These were Arup Nada, Martin Widlake, Heli Helskyaho and Alex Nuijten.

NewImage NewImage

More news coming in 2017.

Blog Posts

One of the things I really enjoy doing is playing with various features of Oracle and then writing some blog posts about them. When writing the books I had to cut back on writing blog posts. I was luck to be part of the 12.2 Database beta this year and over the past few weeks I've been playing with 12.2 in the cloud. I've already written a blog post or two already on this and I also have an OTN article on this coming out soon. There will be more 12.2 analytics related blog posts in 2017.

In 2016 I have written 55 blog posts (including this one). This number is a little bit less when compared with previous years. I'll blame the book writing for this. But more posts are in the works for 2017.

Articles

In 2016 I've written articles for OTN and for Toad World. These included:

OTN
  1. Oracle Advanced Analytics : Kicking the Tires/Tyres
  2. Kicking the Tyres of Oracle Advanced Analytics Option - Using SQL and PL/SQL to Build an Oracle Data Mining Classification Model
  3. Kicking the Tyres of Oracle Advanced Analytics Option - Overview of Oracle Data Miner and Build your First Workflow
  4. Kicking the Tyres of Oracle Advanced Analytics Option - Using SQL to score/label new data using Oracle Data Mining Models
  5. Setting up and configuring RStudio on the Oracle 12.2 Database Cloud Service
ToadWorld
  1. Introduction to Oracle R Enterprise
  2. ORE 1.5 - User Defined R Scripts

Conferences

  1. January - Yes SQL Summit, NoCOUG Winter Conference, Redwood City, CA, USA **
  2. January - BIWA Summit, Oracle HQ, Redwood City, CA, USA **
  3. March - OUG Ireland, Dublin, Ireland
  4. June - KScope, Chicago, USA (3 presentations)
  5. September - Oracle Open World (part of EMEA ACEs session) **
  6. December - UKOUG Tech16 & APPs16

** for these conferences the Oracle ACE Director programme funded the flights and hotels. All other expenses and other conferences I paid for out of my own pocket.

OUG Activities

I'm involved in many different roles in the user group. The UKOUG also covers Ireland (incorporating OUG Ireland), and my activities within the UKOUG included the following during 2016:

  • Editor of Oracle Scene: We produced 4 editions in 2016. Thank you to all who contributed and wrote articles.
  • Created the OUG Ireland Meetup. We had our first meeting in October. Our next meetup will be in January.
  • OUG Ireland Committee member of TECH SIG and BI & BA SIG.
  • Committee member of the OUG Ireland 2 day Conference 2016.
  • Committee member of the OUG Ireland conference 2017.
  • KScope17 committee member for the Data Visualization & Advanced Analytics track.

I'm sure I've forgotten a few things, I usually do. But it gives you a taste of some of what I got up to in 2016.

Monday, December 19, 2016

Auditing Oracle Data Mining model usage

In a previous blog post I talked about how you can rename and comment your Oracle Data Mining models. This is to allow you to easily to see and understand the intended use of the data mining model.

Another feature available to you is to audit the usage of the the data mining models. As your data mining environment grows to many 10s or more typically 100s of models, you will need to have some way of tracking their usage. This can allow you to discover what models are frequently being used and those that are not being used in-frequently. You can then use this information to investigate if there are any issues. Or in some companies I've seen an internal charging scheme in place for each time the models are used.

The following outlines the steps required to setup the auditing of your models and how to inspect the usage.

Note: You will need to the AUDIT_ADMIN role to audit the models.

First create an audit policy for the data mining model in a particular schema.

CREATE AUDIT POLICY oaa_odm_audit_usage 
ACTIONS ALL 
ON MINING MODEL dmuser.high_value_churn_clas_svm;

This creates a policy that monitors all activity on the data mining model HIGH_VALUE_CHURN_CLAS_SVM in the DMUSER schema.

Now we need to enable the policy and allow to to tract all activity on the model.

AUDIT POLICY oaa_odm_audit_usage BY oaa_model_user;

This will track all usage of the data mining model by the schema call OAA_MODEL_USER. We can then use the following query to search for the audit records for the OAA_MODEL_USER schema.

SELECT dbusername,
       action_name, 
       systemm_privilege_used, 
       return_code,
       object_schema, 
       object_name, 
       sql_text
FROM  unified_audit_trail
WHERE object_name = 'HIGH_VALUE_CHURN_CLAS_SVM';

But there is a little problem with using what I've just shown you above. The problem is that it will track all activity on the data mining model. Perhaps this isn't what we really want. Perhaps we only want to track only certain activity of the data mining model. Instead of creating the policy using 'ACTIONS ALL', we can list out the actions or operations we want to track. For example, we want to tract when it is used in a SELECT. The following shows how you can set this up for just SELECT.

CREATE AUDIT POLICY oaa_odm_audit_select 
ACTIONS SELECT 
ON MINING MODEL dmuser.high_value_churn_clas_svm;

AUDIT POLICY oaa_odm_audit_select BY oaa_model_user;

The list of individual audit events you can use include:

  • AUDIT
  • COMMENT
  • GRANT
  • RENAME
  • SELECT

A policy can be setup to tract one or more of these events. For example, if we wanted a policy to track SELECT and GRANT, we would have list each event separated by a comma.

CREATE AUDIT POLICY oaa_odm_audit_select_grant 
ACTIONS SELECT 
ON MINING MODEL dmuser.high_value_churn_clas_svm,
ACTIONS GRANT 
ON MINING MODEL dmuser.high_value_churn_clas_svm,
;

AUDIT POLICY oaa_odm_audit_select_grant BY oaa_model_user;

Monday, December 12, 2016

Renaming & Commenting Oracle Data Mining Models

As your company evolves with their data mining projects, the number of models produced and in use in production will increase dramatically.

Care needs to be taken when it comes to managing these. This includes using meaningful names, adding descriptions of what the model is about or for, and being able to track their usage, etc.

I will look at tracking the usage of the models in another blog post, but the following gives examples of how to rename Oracle Data Mining models and how to add comments or descriptions to these models. This is particularly useful because our data analytics teams have a constant turn over or it has been many months since you last worked on a model and you want a quick idea of what purpose of the model was for.

If you have been using the Oracle Data Mining tool (part of SQL Developer) will will see your model being created with some sort of sequencing numbers. For example for a Support Vector Machine (SVM) model you might see it labelled for classification:

CLAS_SVM_5_22

While you are working on this project you will know and understand what it was about and why it is being used. But afterward you may forget as you will be dealing with many hundreds of models. Yes you could check your documentation for the purpose of this model but that can take some time.

What if you could run a SQL query to find out?

But first we need to rename the model.

DBMS_DATA_MINING.RENAME_MODEL('CLAS_SVM_5_22', 'HIGH_VALUE_CHURN_CLAS_SVM');

Next we will want to add a longer description of what the model is about. We can do this by adding a comment to the model.

COMMENT ON MINING MODEL high_value_churn_clas_svm IS
'Classification Model to Predict High Value Customers most likely to Churn';

We can now see these updated details when we query the Oracle Data Mining models in a user schema.

SELECT model_name, mining_function, algorithm, comments 
FROM user_mining_models;

These are two very useful commands.

Wednesday, December 7, 2016

12.2 DBaaS (Extreme Edition) possible bug/issue with the DB install/setup

A few weeks ago the 12.2 Oracle Database was released on the cloud. I immediately set an account and got my 12.2 DBaaS setup. This was a relatively painless process and quick.

For me I wanted to test out all the new Oracle Advanced Analytics new features and the new features in SQL Developer 4.2 that only become visible when you are using the 12.2 Database.

When you are go to use the Oracle Data Miner (GUI tool) in SQL Developer 4.2, it will check to see if the ODMr repository is installed in the database. If it isn't then you will be promoted for the SYS password.

This is the problem. In previous version of the DBaaS (12.1, etc) this was not an issue.

When you go to create your DBaaS you are asked for a password that will be used for the admin accounts of the database.

But when I entered the password for SYS, I got an error saying invalid password.

After using ssh to create a terminal connection to the DBaaS I was able to to connect to the container using

sqlplus / as sysdba

and also using

sqlplus sys/ as sysdba

Those worked fine. But when I tried to connect to the PDB1 I got the invalid username and/or password error.

sqlplus sys/@pdb1 as sysdba

I reconnected as follows

sqlplus / as sysdba

and then changed the password for SYS with containers=all

This command completed without errors but when I tried using the new password to connect the the PDB1 I got the same error.

After 2 weeks working with Oracle Support they eventually pointed me to the issue of the password file for the PDB1 was missing. They claim this is due to an error when I was creating/installing the database.

But this was a DBaaS and I didn't install the database. This is a problem with how Oracle have configured the installation.

The answer was to create a password file for the PDB1 using the following

% orapwd file=$ORACLE_HOME/dbs/orapwPDB1 password= entries=10

I then changed the password again for SYS, then tried to connect as SYS to the PDB1, and if by magic I was connected.

I then tried installing the ODMr repository again (in SQL Developer) and when I entered the new password for SYS, it worked !

It's a pity that it took Oracle Support 2 weeks to get me to this point.

As 12.2 is a cloud service hopefully Oracle will get that issue fixed soon so that one one else has to suffer like I did.

Monday, December 5, 2016

Evaluating Cluster Dispersion in Oracle Data Mining

When working with the Clustering algorithms, and particularly k-Means, in the Oracle Data Miner tool there is no way of seeing how compact or dispersed the data is within a cluster.

There are a number of measures typically used in various tools and algorithms, but with Oracle Data Miner we are not presented with any of this information.

But if we flip from using the Oracle Data Miner tool to using SQL we can get to see some more details of the clusters produced by the k-Means algorithm along with some additional and useful information.

As I said there are a number of different measures used to evaluate clusters. The one that Oracle uses is called Dispersion. Now there are a few different definitions of what this could be and I haven't been able to locate what is Oracle's own definition of it in any of the documentation.

We can use the Dispersion value as a measure of how compact or how spread out the data is within a cluster. The Dispersion value is a number greater than 0. The lower the value of the more compact the cluster is i.e. the data points are close the the centroid of the cluster. The larger the value the more disperse or spread out the data points are.

The DBMS_DATA_MINING PL/SQL package comes with a function called GET_MODEL_DETAILS_KM. This function returns a record of the form DM_CLUSTERS.

(id                   NUMBER,
 cluster_id           VARCHAR2(4000),
 record_count         NUMBER,
 parent               NUMBER,
 tree_level           NUMBER,
 dispersion           NUMBER,
 split_predicate      DM_PREDICATES,
 child                DM_CHILDREN,
 centroid             DM_CENTROIDS,
 histogram            DM_HISTOGRAMS,
 rule                 DM_RULE)

We can not use the following query to get the Dispersion value for each of the clusters from an ODM cluster model.

SELECT cluster_id,
       record_count,
       parent,
       tree_level,
       dispersion
FROM  table(dbms_data_mining.get_model_details_km('CLUS_KM_3_2'));
NewImage

Tuesday, November 29, 2016

UKOUG Conference 1999 : Who was there & what they were talking about

Yes you read the title of this blog post correctly!

Recently I was doing a bit a clear out and I came across a CD of the UKOUG Conference proceedings from 1999. That was my second UKOUG conference and how times have changed.

NewImage NewImage

The CD contained all the conference proceedings consisting of slides and papers.

Here are some familiar names from back in 1999 who you may find presenting at this years conference, some you might remember as being a regular presenter and some are still presenting but not at this years conference.

  • Jonathan Lewis
  • Carl Dudley
  • Fiona Martin
  • Peter Robson
  • Duncan Mills
  • Kent Graziano
  • John King
  • Toby Price
  • Doug Burns
  • Dan Hotka
  • Joel Goodman

The 1999 Ralph Fiennes did the Keynote speech. I queued up afterwards to get a signed book but they ran out with three people ahead of me :-(

The agenda grid was a bit smaller back then compared to now.

NewImage

I'll see you again in Birmingham this year, in a few days time :-)

I'll be presenting during the Super Sunday and then again on the Tuesday.

When you are at the UKOUG Conference, addend some sessions that are not in your area. You never know what you might learn. Also get out and about, and explore the Christmas market just outside the conference venue.

Monday, November 14, 2016

Using the Identity column for Oracle Data Miner

If you are a user of the Oracle Data Miner tool (the workflow data mining tool that is part of SQL Developer), then you will have noticed that for many of the algorithms you can specify a Case Id attribute along with, say, the target attribute.

NewImage

The idea is that you have one attribute that is a unique identifier for each case record. This may or may not be the case in your data model and you may have a multiple attribute primary key or case record identifier.

But what is the Case Id field used for in Oracle Data Miner?

Based on the documentation this field does not need to have a value. But it is recommended that you do identify an attribute for the Case Id, as this will allow for reproducible results. What this means is that if we run our workflow today and again in a few days time, on the exact same data, we should get the same results. So the Case Id allows this to happen. But how? Well it looks like the attribute used or specified for the Case Id is used as part of the Hashing algorithm to partition the data into a train and test data set, for classification problems.

So if you don't have a single attribute case identifier in your data set, then you need to create one. There are a few options open to you to do this.

  • Create one: write some code that will generate a unique identifier for each of your case records based on some defined rule.
  • Use a sequence: and update the records to use this sequence.
  • Use ROWID: use the unique row identifier value. You can write some code to populate this value into an attribute. Or create a view on the table containing the case records and add a new attribute that will use the ROWID. But if you move the data, then the next time you use the view then you will be getting different ROWIDs and that in turn will mean we may have different case records going into our test and training data sets. So our workflows will generate different results. Not what we want.
  • Use ROWNUM: This is kind of like using the ROWID. Again we can have a view that will select ROWNUM for each record. Again we may have the same issues but if we have our data ordered in a way that ensures we get the records returned in the same order then this approach is OK to use.
  • Use Identity Column: In Oracle 12c we have a new feature called Identify Column. This kind of acts like a sequence but we can defined an attribute in a table to be an Identity Column, and as records are inserted into the the data (in our scenario our case table) then this column will automatically generate a unique number for our data. Again if we need to repopulate the case table, you will need to drop and recreate the table to get the Identity Column to reset, otherwise the newly inserted records will start with the next number of the Identity Column

Here is an example of using the Identity Column in a case table.

CREATE TABLE case_table (
id_column	NUMBER GENERATED ALWAYS AS IDENTITY,
affinity_card 	NUMBER,
age		NUMBER,
cust_gender	VARCHAR2(5),
country_name	VARCHAR2(20)
...
);

You can now use this Identity Column as the Case Id in your Oracle Data Miner workflows.

NewImage

Wednesday, November 9, 2016

New OAA features in Oracle 12.2 Database

The Oracle 12.2c Database has been released and is currently available as a Cloud Service. The on-site version should be with us soon.

A few weeks ago I listed some of the new features that you will find in the Oracle Data Miner GUI tool (check out that blog post). I'll have another blog post soon that looks a bit closer at how the new OAA features are exposed in this tool.

In this blog post I will list most of the new database related features in Oracle 12.2. There is a lot of new features and a lot of updated features. Over the next few months (yes it will take that long) I'll have blog posts on most of these.

The Oracle Advanced Analytics Option new features include:

  • The first new feature is one that you cannot see. Yes that sound a bit odd. But the underlying architecture of OAA has been rebuilt to allow for the algorithms to scale significantly. This is also future proofing OAA for new features coming in future releases of the database.
  • Explicit Semantic Analysis. This is a new algorithm allows us to perform text similarity comparison. This is a great new addition and and much, much easier now compared to what we may have had to do previously.
  • Using R models using SQL. Although we have been able to do this in the previous version of the database, the framework and supports have been extended to allow for greater and easier usage of user defined R scripts and R models with the in-database environment.
  • Partitioned Models. We can now build partitioned mining models. This is where you can specify an attribute and a separate model will be created based on each value in the attribute.
  • Partitioned scoring. Similarly we can now dynamically score the data based on an partition attribute.
  • Extentions to Association Rules. Over the past few releases of the database, additional insights to the workings and decision making of the algorithms have been included. In 12.2 we now have some additional insights for the Association Rules aglorithm where we can now get to see the calculation of values associated with rules.
  • DBMS_DATA_MINING package extended. This PL/SQL package has been extended to include the functionality for the new features listed above. Additional it can now process R algorithms and models.
  • SQL Function changes: Change to the followi ODM related SQL functions to allow for partitioned models. CLUSTER_DETAILS, CLUSTER_DISTANCE, CLUSTER_ID, CLUSTER_PROBABILITY, CLUSTER_SET, FEATURE_COMPARE, FEATURE_DETAILS, FEATURE_ID, FEATURE_SET, FEATURE_VALUE, ORA_DM_PARTITION_NAME, PREDICTION, PREDICTION_BOUNDS, PREDICTION_COST, PREDICTION_DETAILS, PREDICTION_PROBABILITY, PREDICTION_SET
  • New SQL Hint for ODM models. We have had hints in SQL for many, many versions now, but with 12.2c we now have a hint for partitioned models, called GROUPING hint.
  • New CREATE_MODEL function. With the existing CREATE_MODEL function the input data set for the function needed to be defined in a table or accessed using a view. Basically the data needed to resist somewhere. With CREAETE_MODEL2 you can now define the input data set based on a SELECT statement.

In addition to all of these changes there are also some new interesting DB, SQL and PL/SQL new features that are of particular interest for your data science, machine learning, advanced analytics (or whatever the current favourite marketing term is today) projects.

It is going to be a busy few months ahead, working through all of these new features and write blog posts on how to use each of them.

Tuesday, November 1, 2016

Creating and Reading SPSS and SAS data sets in R

NOTE: Several people have contacted me to say that using this R package does not work. The data set 
generated is not readable by SAS. If you encounter this problem then get in touch with the creators 
of Haven for help and support. I'm using R version 3.2.0.
All I an say is, it worked for me!

Have you ever been faced with having to generate a data set in the format that is needed by another analytics tool? or having to generate a data set in a particular format but you don't have the software that generates that format? For example, if you are submitting data to the FDA and other bodies, you may need to submit the data in a SAS formatted file. There are a few ways you can go about this.

One option is that you can use the Haven R package to generate your dataset in SAS and SPSS formats. But you can also read in SAS and SPSS formatted files. I have to deal with these formatted data files all the time, and it can be a challenge, but I've recently come across the Haven R package that has just made my life just a little bit/lots easier. Now I can easily generate SAS and SPSS formatted data sets for my data in my Oracle Database, using R and ORE. ORE we can now use the embedded feature to build the generation of these data sets into some of our end-user applications.

Let us have a look at Haven and what it can do.

Firstly there is very little if any documentation online for it. That is ok so we will have to rely on the documentation that comes with the R packages. Again there isn't much to help and that is because the R package mainly consists of functions to Read in these data sets, functions to Write these data sets and some additional functions for preparing data.

For reading in data sets we have the following functions:

# SAS
read_sas("mtcars.sas7bdat")
# Stata
read_dta("mtcars.dta")
# SPSS
read_sav("mtcars.sav")

For writing data sets we have the following functions:

# SAS
write_sas(mtcars, "mtcars.sas7bdat")
# Stata
write_dta(mtcars, "mtcars.dta")
# SPSS
write_sav(mtcars, "mtcars.sav")

Let us now work through an example of creating a SAS data set. We can use some of the sample data sets that come with the Oracle Database in the SH schema. I'm going to use the data in the CUSTOMER table to create a SAS data set. In the following code I'm using ORE to connect to the database but you can use your preferred method.

> library(ORE)
> # Create your connection to the schema in the DB
> ore.connect(user="sh", password="sh", host="localhost", service_name="PDB12C", 
            port=1521, all=TRUE) 

> dim(CUSTOMERS)
[1] 55500    23
> names(CUSTOMERS)
 [1] "CUST_ID"                "CUST_FIRST_NAME"        "CUST_LAST_NAME"        
 [4] "CUST_GENDER"            "CUST_YEAR_OF_BIRTH"     "CUST_MARITAL_STATUS"   
 [7] "CUST_STREET_ADDRESS"    "CUST_POSTAL_CODE"       "CUST_CITY"             
[10] "CUST_CITY_ID"           "CUST_STATE_PROVINCE"    "CUST_STATE_PROVINCE_ID"
[13] "COUNTRY_ID"             "CUST_MAIN_PHONE_NUMBER" "CUST_INCOME_LEVEL"     
[16] "CUST_CREDIT_LIMIT"      "CUST_EMAIL"             "CUST_TOTAL"            
[19] "CUST_TOTAL_ID"          "CUST_SRC_ID"            "CUST_EFF_FROM"         
[22] "CUST_EFF_TO"            "CUST_VALID"      

Next we can prepare the data, take a subset of the data, reformat the data, etc. For me I just want to use the data as it is. All I need to do now is to pull the data from the database to my local R environment.

dat <- ore.pull(CUSTOMERS)

Then I need to load the Haven library and then create the SAS formatted file.

library(haven)
write_sas(dat, "c:/app/my_customers.sas7bdat")

That's it. Nice and simple.

But has it worked? Has it created the file correctly? Will it load into my SAS tool?

There is only one way to test this and that is to only it in SAS. I have an account on SAS OnDemand with access to several SAS products. I'm going to use SAS Studio.

Well it works! The following image shows SAS Studio after I had loaded the data set with the variables and data shown.

NewImage

WARNING: When you load the data set into SAS you may get a warning message saying that it isn't a SAS data set. What this means is that it is not a data set generated by SAS. But as you can see in the image above all the data got loaded OK and you can work away with it as normal in your SAS tools.

The next step is to test the loading of a SAS data set into R. I'm going to use one of the standard SAS data sets called PVA97NK.SAS7BDAT. If you have worked with SAS products then you will have come across this data set.

When you use Haven to load in your SAS data set, it will create the data in tribble format. This is a slight varient of a data.frame. So if you want the typical format of a data.frmae then you will need to convert the loaded data, as shown in the following code.

> data_read <- read_sas("c:/app/pva97nk.sas7bdat")
> dim(data_read)
[1] 9686   28
> d<-data.frame(data_read)
> class(data_read)
[1] "tbl_df"     "tbl"        "data.frame"
> class(d)
[1] "data.frame"
> head(d)
  TARGET_B       ID TARGET_D GiftCnt36 GiftCntAll GiftCntCard36 GiftCntCardAll
1        0 00014974       NA         2          4             1              3
2        0 00006294       NA         1          8             0              3
3        1 00046110        4         6         41             3             20
...

I think this package to going to make my life a little bit easier, and if you work with SPSS and SAS data sets then hopefully some of your tasks have become a little bit easier too.

Tuesday, October 25, 2016

Oracle Text, Oracle R Enterprise and Oracle Data Mining - Part 5

In this 5th blog post in my series on using the capabilities of Oracle Text, Oracle R Enterprise and Oracle Data Mining to process documents and text, I will have a look at some of the machine learning features of Oracle Text.

Oracle Text comes with a number of machine learning algorithms. These can be divided into two types. The first is called 'Supervised Learning' where we have two machine learning algorithms for classification type of problem. The second type is called 'Unsupervised Learning' where we have the ability to use clustering machine learning algorithms to look for patterns in our text documents and to find similarities between documents based on their contents.

It is this second type of document clustering that I will work through in this blog post.

When using clustering with text documents, the machine learning algorithm will look for patterns that are common between the documents. These patterns will include the words used, the frequency of the words, the position or ordering of these words, the co-occurance of words, etc. Yes this is a large an complex task and that is why we need a machine learning algorithm to help us.

With Oracle Text we only have one clustering machine learning algorithm available to use. When we move onto using the Oracle Advanced Analytics Option (Oracle Data Mining and Oracle R Enterprise) we more algorithms available to us.

With Oracle Text the clustering algorithm is called k-Means. In a way the actual algorithm is unimportant as it is the only one available to us when using Oracle Text. To use this algorithm we have the CTX_CLS.CLUSTERING procedure. This procedure takes the documents we want to compare and will then identify the clusters (using hierarchical clustering) and will then tells us, for each document, what clusters the documents belong to and they probability value. With clustering a document (or a record) can belong to many clusters. Typically in the text books we see clusters that are very distinct and are clearly separated from each other. When you work on real data this is never the case. We will have many over lapping clusters and a data point/record can belong to one or more clusters. This is why we need the probability vale. We can use this to determine what cluster our record belongs to most and what other clusters it is associated with.

Using the example documents that I have been using during this series of blog posts we can use the CTX_CLS.CLUSTERING algorithm to cluster and identify similarities in these documents.

We need to setup the parameters that will be used by the CTX_CLS.CLUSTERING procedure. Tell it to use the k-Means algorithm and then the number of clusters to generate. As with all Oracle Text procedures or algorithms there are a number of settings you can configure or you can just accept the default values.

exec ctx_ddl.drop_preference('Cluster_My_Documents');
exec ctx_ddl.create_preference('Cluster_My_Documents','KMEAN_CLUSTERING');
exec ctx_ddl.set_attribute('Cluster_My_Documents','CLUSTER_NUM','3');

The code above is an example of the basics of what you need to setup for clustering. Other attribute or cluster parameter setting available to you include, MAX_DOCTERMS, MAX_FEATURES, THEME_ON, TOKEN_ON, STEM_ON, MEMORY_SIZE and SECTION_WEIGHT.

Now we can run the CTX_CLS.CLUSTERING procedure on our documents. This procedure has the following parameters.

- The Oracle Text Index Name

- Document Id Column Name

- Document Assignment (cluster assignment) Table Name. This table will be created if it doesn't already exist

- Cluster Description Table Name. This table will be created if it doesn't already exist.

- Name of the Oracle Text Preference (list)

exec ctx_cls.clustering(
'MY_DOCUMENTS_OT_IDX',
'DOC_PK',
'OT_CLUSTER_RESULTS',
'DOC_CLUSTER_DETAILS',
'Cluster_My_Documents');

When the procedure has completed we can now examine the OT_CLUSTER_RESULTS and the DOC_CLUSTER_DETAILS tables. The first of these (OT_CLUSTER_RESULTS) allows us to see what documents have been clustered together. The following is what was produced for my documents.

SELECT d.doc_pk, 
       d.doc_title, 
       r.clusterid, 
       r.score 
FROM my_documents d, 
     ot_cluster_results r 
WHERE d.doc_pk = r.docid;

NewImage

We can see that two of the documents have been grouped into the same cluster (ClusterId=2). If you have a look back at what these documents are about then you can see that yes these are very similar. For the other two documents we can see that they have been clustered into separate clusters (ClusterId=4 & 5). The clustering algorithms have said that they are different types of documents. Again when you examine these documents you will see that they are talking about different topics. So the clustering process worked !

You can also explore the various features of the clusters by looking that he DOC_CLUSTER_DETAILS table. Although the details in this table are not overly useful but it will give you some insight into what clusters the k-Means algorithm has produced.

Hopefully I've shown you how easy it is to setup and use the clustering feature of Oracle Text.

WARNING: Before using the Clustering or Classification with Oracle Text, you need to check with your local Oracle Sales representative about if there is licence implication. There seems to be some mentions the the algorithms used are those that come with Oracle Data Mining. Oracle Data Mining is a licence cost option for the database. So make sure you check before you go using these features.

Saturday, October 22, 2016

Our first OUG Ireland Meet-up

Last Thursday evening (20th October) we had our first Meet-up event for OUG Ireland.

Up to recently we have had a one or two full day SIG events that covered both the Tech and BA/Big Data areas. But we have been finding it increasingly difficult to get speakers and attendees to take a full day out of work to attend the SIG events. This was particularly true for a SIG event we had scheduled to happen in early October. But by the end of August things were not coming together for us, so it was time to try something new.

Over the past couple of years Meet-ups have been growing in numbers and in popularity. We (the OUG Ireland SIG committee) have been keeping an eye on this in the UK and Ireland.

It was time to get this concept a try.

What did it entail and what venue did you use?

The first thing we needed to do was to arrange a venue. A very popular location for Meet-ups in Ireland is in one of Bank of Ireland branches. This is Bank of Ireland on Grand Canal Dock in Dublin. It is one of their enterprise centres and is open during the day for meetings, as a workspace and it also operates as a branch. In the evenings and on a Saturday morning it is available for groups to hold meetings for larger groups and Meet-ups. We have the venue from 6pm-8pm.

What presentations did you have?

After securing a venue the we then decided to have the theme of the Meet-up to be about 'Updates from Oracle Open World'. Most of the SIG committee was at Oracle Open World, so that should be easy enough to put a few presentations together and we got the local Oracle office to joins us to.

NewImage

What about catering?

The venue very kindly supplies some soft drinks, tea, coffee, a few beers, along with some sandwiches and pastries. All for free!

So far we have a free venue, free catering and the committee for presenters.

NewImage

How did you advertise the event?

The only other thing we needed to do now was to advertise the event. For this we used a combination of EventBrite and Meet-up.com for this, along with our own contacts. Plus some of our friends helped to spread the word. This worked really well. We ended up getting roughly the same number of people registering for the event on each platform. We had 98 registrations on these websites.

Something we were warned about is that a lot of people will register, but if you get 40% of those turning up for the event then you are doing well. We got 48 attendees (=50%). We were delighted with this. For our full day SIG events we might have had 20-25 attendees.

How much this this event cost?

It cost us zero euro/dollars/sterling. As there was no admin, advertising costs, catering, room hire, nothing. Well that is not entirely true. There was a small cost and that was for our membership fee to be on Meet-up. That cost me about $30 for 6 months (unlimited plan). Yes I've paid that myself.

What was the feedback after the event?

The feedback was fantastic. People loved the new format, loved that it was in the evening after work, liked the short length presentations, liked that it was free, etc.

I also asked people if they might be interested in presenting at a future Meet-up. Personally I had 5 people talk to me about this. The over committee members also had people talk to them about it. It seems people are interested in trying the shorter format presentations, as it is not as daunting as presenting at a conference. A conference seems to be more formal and a step up in presenting levels.

NewImage NewImage

What next?

Well we have the same venue booked for 12th January and 11th May. We have our 2 presenters for the 12th January already. We actually had those before our first meet-up. Plus for the 11th May we have some possible presenters and I just need to work with them to see who will get on the agenda.

The plan was to have 3-4 of these each year. Based on the feedback and the level of interest we might need to have a few more. But it is still early days and we need to see how things develop.

Our Meet-up was in Dublin. Ideally we want to bring this to over regions. For example we could do the same in Belfast and Cork, and possibly Limerick and Galway. This is something we are looking at and in 2017 we will definitely have a Meet-up in one (or two) of those locations. That would bring us up to 4-5 Meet-ups in 2017.

Thank you to everyone who attended and everyone how helped to make this happen.