Monday, May 20, 2024
HomeMatlabOverstock no extra! – Hear from the winners of MATHack Western Cape...

Overstock no extra! – Hear from the winners of MATHack Western Cape 2022 » Scholar Lounge


Becoming a member of us immediately is a bunch of scholars from Stellenbosch College in South Africa. Bianca Harber, Gabrielle Liebenberg, and Rosanne Maritz received first place at MATHack Western Cape 2022 by constructing a machine studying primarily based software to assist small companies of their group take care of overstocking. Bianca, Gabrielle and Rosanne, over to you…

Inspiration

We had been offered with the duty of discovering a inventive manner to make use of MATLAB to assist small companies inside the Western Cape through the COVID-19 pandemic. Throughout the lockdowns in South Africa, one of many many issues small companies confronted was what to do with extra inventory. We determined to handle the basis of this concern: why is there extra inventory to start with? This drawback impressed us to create a instrument for small companies that can take a number of information options into consideration from a dataset of previous gross sales to precisely estimate gross sales sooner or later primarily based on domestically related parameters.

Breaking down the issue

With all of the advances within the area, we knew it could be attainable to simply make this prediction instrument with the usage of information science and machine studying strategies and algorithms. It was solely a matter of discovering the precise strategies for our particular drawback.

We knew that we wanted to implement a regression algorithm to acquire the anticipated output. We chosen common month-to-month climate, value, and lockdown stage as our machine studying options, as we thought these would have the largest impact on gross sales.

How did we implement it?

We utilised MATLAB’s Regression Learner app to to coach a number of regression fashions, take a look at pattern information in opposition to these fashions, after which we chosen the mannequin that yielded probably the most correct predictions to make use of in our app. An choice so as to add complete gross sales per product monthly, common month temperature and present lockdown stage was added to permit the consumer to generate predictions.

Outcomes

We delivered a minimal viable product that will have the ability to:

  • Import collected information, similar to that from a point-of-sale system
  • Apply the machine studying mannequin on the information
  • Predict future gross sales primarily based on recognized developments
  • Permit the consumer to view these gross sales predictions by product

The interface for our product, which was simply designed with MATLAB App Designer, is proven beneath:

Picture2.png

The applied mannequin made sound and correct predictions, with only some factors of information to work with.

Picture1.png

Key Takeaways

Some potential expansions of the product that we thought can be cool to implement had been:

  • The usage of location information (which the MATLAB Cellular software can be utilized for) to routinely load climate information into the system
  • The flexibility to have this software work on all cell platforms
  • The storage of information in an internet database for cross-device synchronization and prevention of information loss
  • Automated inventory monitoring

By our expertise on this hackathon, we actually noticed the facility of MATLAB and its wide selection of purposes in trade. It was inspiring to see first-hand that you just don’t should be a machine studying skilled to implement this instrument in an issue you are attempting to unravel. It was an ideal alternative to check our drawback fixing and inventive design abilities whereas making an attempt to unravel an issue confronted by actual folks.



RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Most Popular

Recent Comments