Wednesday, March 13, 2019

Selecting Optimal Parameters for XGBoost Model Training

There is always a bit of luck involved when selecting parameters for Machine Learning model training. Lately, I work with gradient boosted trees and XGBoost in particular. We are using XGBoost in the enterprise to automate repetitive human tasks. While training ML models with XGBoost, I created a pattern to choose parameters, which helps me to build new models quicker. I will share it in this post, hopefully you will find it useful too.

I’m using Pima Indians Diabetes Database for the training, CSV data can be downloaded from here.

This is the Python code which runs XGBoost training step and builds a model. Training is executed by passing pairs of train/test data, this helps to evaluate training quality ad-hoc during model construction:

Key parameters in XGBoost (the ones which would affect model quality greatly), assuming you already selected max_depth (more complex classification task, deeper the tree), subsample (equal to evaluation data percentage), objective (classification algorithm):
  • n_estimators — the number of runs XGBoost will try to learn 
  • learning_rate — learning speed 
  • early_stopping_rounds — overfitting prevention, stop early if no improvement in learning 
When is executed with verbose=True, you will see each training run evaluation quality printed out. At the end of the log, you should see which iteration was selected as the best one. It might be the number of training rounds is not enough to detect the best iteration, then XGBoost will select the last iteration to build the model.

With matpotlib library we can plot training results for each run (from XGBoost output). This helps to understand if iteration which was chosen to build the model was the best one possible. Here we are using sklearn library to evaluate model accuracy and then plotting training results with matpotlib:

Let’s describe my approach to select parameters (n_estimators, learning_rate, early_stopping_rounds) for XGBoost training.

Step 1. Start with what you feel works best based on your experience or what makes sense
  • n_estimators = 300 
  • learning_rate = 0.01 
  • early_stopping_rounds = 10 
  • Stop iteration = 237 
  • Accuracy = 78.35% 
Results plot:

With the first attempt, we already get good results for Pima Indians Diabetes dataset. Training was stopped at iteration 237. Classification error plot shows a lower error rate around iteration 237. This means learning rate 0.01 is suitable for this dataset and early stopping of 10 iterations (if the result doesn’t improve in the next 10 iterations) works.

Step 2. Experiment with learning rate, try to set a smaller learning rate parameter and increase number of learning iterations
  • n_estimators = 500 
  • learning_rate = 0.001 
  • early_stopping_rounds = 10 
  • Stop iteration = didn’t stop, spent all 500 iterations 
  • Accuracy = 77.56% 
Results plot:

Smaller learning rate wasn’t working for this dataset. Classification error almost doesn’t change and XGBoost log loss doesn’t stabilize even with 500 iterations.

Step 3. Try to increase the learning rate.
  • n_estimators = 300 
  • learning_rate = 0.1 
  • early_stopping_rounds = 10 
  • Stop iteration = 27 
  • Accuracy = 76.77% 
Results plot:

With increased learning rate, the algorithm learns quicker, it stops already at iteration Nr. 27. XGBoost log loss error is stabilizing, but the overall classification accuracy is not ideal.

Step 4. Select optimal learning rate from the first step and increase early stopping (to give the algorithm more chances to find a better result).
  • n_estimators = 300 
  • learning_rate = 0.01 
  • early_stopping_rounds = 15 
  • Stop iteration = 265 
  • Accuracy = 78.74% 
Results plot:

A slightly better result is produced with 78.74% accuracy — this is visible in the classification error plot.


Wednesday, March 6, 2019

Prepare Your Data for Machine Learning Training

The process to prepare data for Machine Learning model training to me looks somewhat similar to the process of preparing food ingredients to cook dinner. You know in both cases it takes time, but then you are rewarded with tasty dinner or a great ML model.

I will not be diving here into data science subject and discussing how to structure and transform data. It all depends on the use case and there are so many ways to reformat data to get the most out of it. I will rather focus on simple, but a practical example — how to split data into training and test datasets with Python.

Make sure to check my previous post, today example is based on a notebook from this post — Jupyter Notebook — Forget CSV, fetch data from DB with Python. It is explained there how to load data from DB and construct a data frame.

This Python code snippet builds train/test datasets:

The first thing is to assign X and Y. Data columns assigned to X array are the ones which produce decision encoded in Y array. We assign X and Y by extracting columns from the data frame.

In the next step train X/Y and test X/Y sets are constructed by function train_test_split from sklearn module. You must import this function in Python script:

from sklearn.model_selection import train_test_split

One of the parameters for train_test_split function — test_size. This parameter controls the proportion of test data set size taken from the entire data set (~30% in this example).

Parameter stratify is enforcing equal distribution of Y data across train and test data sets.

Parameter random_state ensures data split will be the same in the next run too. To change the split, it is enough to change this parameter value.

Function train_test_split returns four arrays. Train X/Y and test X/Y pairs can be used for train and test ML model. Data set shape and structure can be printed out too for the convenience purpose.

Sample Jupyter notebook available on GitHub. Sample credentials JSON file.

Saturday, February 23, 2019

Oracle JET Table with Template Slots for Custom Cells

Oracle JET table comes with template slot option. This is helpful to build generic functionality to render custom cell within the table.

In this example, custom cells are used to render dates, amount and risk gauge:

While implementing Oracle JET table it is a best practice to read table column structure from a variable, not to define the entire structure in HTML itself. Property columns refer to the variable. Template called cellTemplate is a default template to render cell content:

Table column structure is defined in JS. To apply specific cell template, it is specified in column definition:

Table data is static in this example and coming through JSON array based on JET Array Data Provider:

Sample code is available on GitHub.

Wednesday, February 20, 2019

Intercepting ADF Table Column Show/Hide Event with Custom Change Manager Class

Ever wondered how to intercept ADF table column show/hide event from ADF Panel Collection component? Yes, you could use ADF MDS functionality to store user preference for table visible columns. But what if you would want to implement it yourself without using MDS? Actually, this is possible through custom persistence manager class. I will show you how.

If you don't know what I'm talking about. Check below screenshot, this popup comes out of the box with ADF Panel Collection and it helps to manage table visible columns. Pretty much useful, especially for large tables:

Obviously, we would like to store user preference and next time the user comes back to the form, he should see previously stored setup for the table columns. One way to achieve this is to use out of the box ADF MDS functionality. But what if you don't want to use it? Still possible - we can catch all changes done through Manage Columns popup in custom Change Manager class. Extend from SessionChangeManager and override only a single method - addComponentChange. This is the place where we intercept changes and could log them to DB for example (later on form load, we could read table setup and apply it before fragment is rendered):

Register custom Change Manager class in web.xml:

Manage Columns popup is out of the box functionality offered by ADF Panel Collection component:

Method addComponentChange will be automatically invoked and you should see similar output when changing table columns visibility:

Download sample application code from my GitHub repository.

Friday, February 15, 2019

ADF Performance Improvement with Nginx Compression

We are using Nginx web server for Oracle ADF WorkBetter hosted demo hosted on DigitalOcean cloud server. Nginx helps to serve web application content fast and offer improved performance. One of the important tuning options - content compression, Nginx does this job well and is simple to setup.

Content compression doesn't provide direct runtime performance, a browser would run the same code, doesn't matter it was compressed or not. But it brings improved perceived performance (which is very important), network time is way faster, because of reduced content size. Oracle ADF is a server-side framework, each request would bring content from the server - faster this content comes, means better application performance.

1. Content Compression = OFF

Let see stats, when no content compression applied (using our Oracle ADF WorkBetter hosted demo).

Page load size is 2.69 MB transferred. Finish time 1.55 s:

Navigation to the employee section generates 165.76 KB and finish time 924 ms:

Navigation to employee compensation generates 46.19 KB and finish time 494 ms:

2. Nginx compression

Compression is simple to setup in Nginx. Gzip settings are set in nginx.conf, make sure to list all content types which must be supported for compression. Restart nginx process after new settings are saved in nginx.conf:

3. Content Compression = ON

Page load size is 733.84 KB transferred. Finish time 1.48 s:

Navigation to the employee section generates 72.75 KB and finish time 917 ms:

Navigation to employee compensation generates 7.59 KB and finish time 498 ms:

Monday, February 11, 2019

Jupyter Notebook — Forget CSV, fetch data from DB with Python

If you read a book, article or blog about Machine Learning — high chances it will use training data from CSV file. Nothing wrong with CSV, but let’s think if it is really practical. Wouldn’t be better to read data directly from the DB? Often you can’t feed business data directly into ML training, it needs pre-processing — changing categorial data, calculating new data features, etc. Data preparation/transformation step can be done quite easily with SQL while fetching original business data. Another advantage of reading data directly from DB — when data changes, it is easier to automate ML model re-train process.

In this post I describe how to call Oracle DB from Jupyter notebook Python code.

Step 1 

Install cx_Oracle Python module:

python -m pip install cx_Oracle

This module helps to connect to Oracle DB from Python.

Step 2

cx_Oracle enables to execute SQL call from Python code. But to be able to call remote DB from Python script, we need to install and configure Oracle Instant Client on the machine where Python runs.

If you are using Ubuntu, install alien:

sudo apt-get update 
sudo apt-get install alien 

Download RPM files for Oracle Instant Client and install with alien:

alien -i oracle-instantclient18.3-basiclite-–1.x86_64.rpm 
alien -i oracle-instantclient18.3-sqlplus-–1.x86_64.rpm 
alien -i oracle-instantclient18.3-devel-–1.x86_64.rpm 

Add environment variables:

export ORACLE_HOME=/usr/lib/oracle/18.3/client64 

Read more here.

Step 3 

Install Magic SQL Python modules:

pip install jupyter-sql 
pip install ipython-sql 

Installation and configuration complete.

For today sample I’m using Pima Indians Diabetes Database. CSV data can be downloaded from here. I uploaded CSV data into the database table and will be fetching it through SQL directly in Jupyter notebook.

First of all, the connection is established to the DB and then SQL query is executed. Query result set is stored in a variable called result. Do you see %%sql — this magic SQL:

Username and password must be specified while establishing a connection. To avoid sharing a password, make sure to read password value from the external source (it could be simple JSON file as in this example or more advanced encoded token from keyring).

The beauty of this approach — data fetched through SQL query is out of the box available in Data Frame. Machine Learning engineer can work with the data in the same way as it would be loaded through CSV:

Sample Jupyter notebook available on GitHub. Sample credentials JSON file.

Tuesday, February 5, 2019

JDeveloper 12c IDE Performance Boost

There is a way to optimize JDeveloper 12c IDE performance by disabling some of the features you are not using.

I was positively surprised with improved JDeveloper responsiveness after turning off some of the features. ADF BC, Task Flow, and ADF Faces wizards started to respond in a noticeably faster way. Simple change and big performance gain, awesome.

One of the strongest JDeveloper performance improvements come from disabling TopLink feature. Ironically - TopLink is an abandoned product (12.1.3 was the last release). I remember back in 2006 TopLink was very promising and it was almost becoming the default platform for ADF Model. One of the old blog posts written by me related to TopLink - External Transaction Service in Oracle TopLink. But luckily it was overshadowed by ADF BC.

These are the features I disabled in my JDeveloper to get performance gain:

Monday, February 4, 2019

Cross Field Form Validation in Oracle JET

JET keeps evolving and in the latest versions  - toolkit provides improved support for form cross-field validation. It is much easier to implement validation than it was before. I will show it in this example.

Example of the data entry form. Validation logic:

- Invoice Date before Payment Due Date and Payment Date
- Payment Due Date before Payment Date

Example when two fields fail validation:

JET provides component called validation group. Form can be wrapped by this component to identify if any validation errors are reported there. For example, when calling JS function, before proceeding with the function code - we can check if validation group contains errors:

Input field can be assigned with custom validator function:

Example of validation function code where cross-field validation logic is implemented - we compare field value with other fields. If validation rule condition is false - validation error is thrown:

Example of function code, where validation group is checked for errors. If there are errors in the current validation group - errors are displayed and the first field with error is focused:

Download sample code from my GitHub repo.

Saturday, January 26, 2019

Search Form in Oracle Visual Builder based on ADF BC REST

Oracle Visual Builder supports ADF BC REST out of the box. Build service connection using "Define by Specification" wizard:

Wizards support ADF as API type. Add describe at the end of the REST URL, this will bring metadata for exposed ADF BC REST service (information about attribute types, etc.):

List of endpoints will be populated automatically. You could select all endpoints to be supported for your connection or select only few:

The most typical thing you would do with endpoint - map it with the table to display collection data. You would drag and drop Oracle JET table into VBCS page and choose Add Data option to map it with the service connection:

In the wizard you would select previously defined service connection:

There is a way to switch wizard to detailed view and choose from multiple endpoints available for the connection:

In the next step, you would select service attributes to be displayed in table columns. All declarative, sweet:

In Visual Builder at any point you can quickly test application, it will load in separate browser tab (or you could switch app to Live mode and test page functionality directly in VBCS window):

Every action in Visual Builder is handled through events. For example, this event is mapped with Reset button (you can see it in structure tab on the left):

At any point, you can switch to source view and check (or edit) HTML/JET code which is generated for you by Visual Builder. So cool, imagine typing and copy-pasting all this text by hand, tough and time-consuming (you could do better things in your life than copy-pasting HTML code):

Let's explain how search form logic is done in this sample. I have defined page scope variable type, this type would hold search attribute name, type and operation:

Create as many variables based on this type, as many search criteria items you will expect to have. Make sure to provide attribute and operation names (leave value property empty, this will be assigned by user):

Map search form fields with variables:

Create an event for Search button, which calls search action chain:

In action chain we can define search logic. Before executing search criteria, we need to prepare search criteria array (normally this step could be skipped, but there is issue in current Visual Builder, it fails to execute criteria search, when at least one of the criteria items empty). Calling custom JavaScript function where search criteria array will be prepared:

Custom JavaScript function, it helps to prepare array to be based to criteria (if search item is not set, we are assigning empty value):

Result of the function is mapped with service connection criteria, search will be executed automatically:

Table pagination is handled automatically too. Make sure to specify scroll policy = loadMoreOnScroll and define fetch size:


1. Sample source code on my GitHub
2. Blog from Shay - Filtering Data Providers with Compound Conditions in Visual Builder
3. Blog from Shay - Oracle JET UI on Top of Oracle ADF With Visual Builder
4. My previous post about query logic in Visual Builder - Query Logic Implementation in VBCS for ADF BC REST

Monday, January 21, 2019

Announcing Hosting for Oracle ADF Rich Client and Oracle ADF WorkBetter Demos

If you are curious about how Oracle ADF works or want to explore a rich set of ADF Faces components - welcome to access Oracle ADF demo apps hosted on our cloud server.

We launched a dedicated website Oracle ADF Components. Hosted demos:

1. ADF Faces Rich Client
2. ADF Work Better

These demo apps can be downloaded from Oracle, you could run them on your own environment too. But sometimes it is useful to have apps online for quick access.

Saturday, January 19, 2019

Oracle ADF BC Reusing SQL from Statement Cache

Oracle ADF BC by default is trying to reuse prepared SQL query from statement cache. It works this way when ADF BC runs with DB pooling off (jbo.doconnectionpooling=false). Normally we tune ADF application to run with DB pooling on (jbo.doconnectionpooling=true), this allows to release unused DB connection back to the pool when a request is completed (and in this case, statement cache will not be used anyway). If View Object is re-executed multiple times during the same request - in this situation, it will use statement cache too.

However, there are cases when for specific View Object you would want to turn off statement cache usage. There could be multiple reasons for this - for example, you are getting Closed Statement error after it tries to execute SQL for statement obtained from statement cache. Normally you would be fine using statement cache, but as I said - there are those special cases.

We are lucky because there is a way to override statement cache usage behavior. This can be done in View Object implementation class either for particular View Object or in the generic class.

After View Object was executed, check the log. If this is not the first execution, you will see log message - "reusing defined prepared statement". This means SQL will be reused from statement cache:

To control this behavior, override getPreparedStatement method:

We create new prepared statement in this method, instead of reusing one from the cache.

As a result - each time View Object is executed, there is no statement cache usage:

Download sample application from GitHub repo.

Saturday, January 12, 2019

On-Premise Machine Learning with XGBoost (Katana 19.1)

Happy to announce Katana 19.1 release with complete on-premise support for Machine Learning.

You can run Machine Learning (ML) models on Cloud (Amazon SageMaker, Google Cloud Machine Learning, etc.). I believe it is important to understand how to run Machine Learning in your own environment too. Without this knowledge ML skills set would not be complete. There are multiple reasons for this. Not everyone is using Cloud and you must provide on-premise solution. Without getting your hands dirty and configuring environment yourself, you would miss an exciting opportunity to learn more about ML.

Read more here.