How to Integrate Machine Learning into Microsoft Access Applications

In this episode, we're going to level up our Microsoft Access application by integrating Machine Learning into it, so that we can see classification of customers that will show predictions of a fictional bank scenario where customers become part of a "preferred program" or not. We'll integrate the work we did in our last video where we created an Artificial Neural Network (ANN) to do this task. By integrating powerful Python scripting with our application, we will harness the power of Machine Learning in our MS Access application.
I have been asked many times over the years on how to implement weighting, scoring, and predictive systems in complex environments. This episode's ANN integration can be used in a wide range of contexts where MS Access is used, and though the actual ML model used is one of the simplest that can be used, it is still FAR more powerful than most other weighting and scoring systems that can be (easily) made with MS Access. And, when you consider that the model can be extended into Deep Learning territory to tackle greater complexity with more precision, you can see the power that can be added to any of your applications.
Related Videos:
How to Use a File Dialog in MS Access to Allow Users to Pick a File or Folder
• How to Use a File Dial...
How to Write a Custom Text File Using MS Access VBA
• How to Write a Custom ...
How to Integrate Machine Learning into Microsoft Access Applications
You are watching this video now!
How to Use Base64 in Microsoft Access
• How to Use Base64 in M...
How to Use Sleep to Pause VBA in MS Access
• How to Use Sleep to Pa...
How to Open Excel from Access Using VBA
• How to Open Excel from...
How to Use vbScript to Open and Work with Data in Access accdb Files - Late Binding Example
• How to Use vbScript to...
Some things to remember when watching this video:
- The ANN we created in my last video then integrated here is a simple one. It used many defaults (ie. learning rate, batch size)
- It could easily be tuned for higher accuracy.
- The purpose of this video was to demonstrate a sample integration. Every integration and application will be different.
- Many datasets easily lend themselves to classification, but many do not and it is hard to get accurate classification.
- This technique can also be used for computer vision (photo classification)
- This example is called Supervised Learning, where we provide the labels of each case for the computer to learn from.
Join me on Patreon!
/ mackenziedataengineering
Demo of my BZ RDP Cloaker:
www.patreon.co...
Want the code from this video?
mackenziemacken...
Interested in transforming your career or finding your next gig?
system.billzon...
Want my team to do a project for you? Let's get to it!
system.billzon...
Contact form you can find at www.mackenziema...
Follow me on social media:
/ mackenziedataanalytics
/ seamacke
/ seamacke
/ seamacke
/ psmackenzie
Get Microsoft Office including Access:
click.linksyne...
Got a KZbin Channel? I use TubeBuddy, it is awesome. Give it a try:
www.tubebuddy....
#msaccess #datascience #python

[@interestingamerican3100]

Holy shit! Probably a KZbin first to combine "Machine Learning" and "Microsoft Access".

[@seanmackenziedataengineering]

I thought this would be a good one. Lots of people ask me about using different weighted schemes to classify things in their database application. I thought why not bring a sledgehammer to this issue. Access has the power to utilize other technologies. Let’s add artificial intelligence to our applications!

[@alaanasser3243]

So glad to have found this channel . I have been asking on many forms that predictions with ML are good and well but how to actually operationalize the outputs. This video is one of a kind .. Many thanks, Sir!

[@seanmackenziedataengineering]

Welcome aboard! Thanks for the feedback!

[@RS-tx4bu]

Sean, thank you so much for making this video. This is a revolutionary video that can forever change the future for many. It is like one of those white papers that the great minds wrote in history. You have just given RAD GUI power to Python developers. And, at the same time, you have given the power of Python to all VBA And MS Office suite devs. This is hands down one of the best videos I have ever seen. Literally, from this point forward, the sky, and one's imagination, is the limit! As always, keep up The enlightening work.

[@seanmackenziedataengineering]

Many thanks, and I’m glad you enjoyed this video! I’m always interested in useful integrations. More to come.

[@siclucealucks]

This is one of the videos where you start watching 1 min and you already want to give a like :D

[@seanmackenziedataengineering]

Awesome! Smash that Like button!!! 👍

[@siclucealucks]

@@seanmackenziedataengineering surely i did ;)

[@seanmackenziedataengineering]

Also watch my previous video for the script we integrated in this video: kzbin.info/www/bejne/hXvCpH6gbMykj6M if you want to learn how to apply a simple 4 column (or "feature") ANN with 64 neurons that produces a two column output with ID and prediction for a binary prediction (ie. 0 or 1). Do you have a cool situation where you think A.I. would be awesome in your database? Need the files used in this video? Check out my downloads page: mackenziemackenzie.com/downloads

[@raziq300]

Sir kindly make some tutorial on insert record into tables from two or more than two relational from to tables

[@seanmackenziedataengineering]

I think this could be a good idea for a video. Can you describe it more?

[@raziq300]

@@seanmackenziedataengineering sir i want to insert record from Customer Invoice with Unbound text fields into Table

[@raziq300]

@@seanmackenziedataengineering Sir also want to block or deny the autosave option in Ms access.

[@lucassouto4902]

great class. I need to know How to create a database and approval form for service requests?

[@seanmackenziedataengineering]

You might be able to find one in the "templates" that you see when you open Access. Try a few - if there is one close to what you need, you can modify it and try using it. Or, just try building one yourself!

[@lucassouto4902]

@@seanmackenziedataengineering found a model. thanks