Skip to main content

Introduction

Vector AI modul

This repository contains the basic framework solutions for each vector AI solutions.

Getting started

Basic informations

Machine Learning solutions are always based on the learning data. From a training dataset we will be developing a machine learning model, which will be trained on our training dataset. Hence it is not a classis programming experience, since these models only gives predictions on future new data, and its accuracy may change over time due to many reasons, for example data drifts.
Because of this, when developing a machine learning model, one should do that individually with every customer. Because it is completly dependent on the data, a model result may vary per customers. So in the repository we will be develeping demo python applications for each goal, and these should be tailored by customers from their data. The goal of these repositories, is not to start from sracth with each project.
Beacause of the above, every Machine Learning AI solution should be considered a R&D project.

Education

Basic Machine Learning related education can be found here. Contains python codes with lot of examples.

tsbroker\oktatas\Cubix Machine Learning Engineer

In-work

  1. Recommendation systems
  • user based content filtering
  • item based content filtering
  • apriori
  1. Clustering
  • Customer clustering based on purchasing volumes
  • Product clustering based on product properties -> similiar products
  • Customer churn detection

Road-map

  1. Time series - Stock predictions
  2. Anomaly detection - user behaviour
  3. Anomaly detection - unusual prices
  4. OCR - invoice pdf recognition
Back to top