We offer a quickscan based on different Machine Learning tools for optimizing & predictive modelling. Based on specific problems with regard to new business or marketing questions we focus on possible blind spots that we can solve through machine learning analysis. We use a 5 step approach for a succesful quickscan (see also blog 1 A 5-Step Process to Get More Out of Your Organization’s Data. First we check all available and missing data. An important step is to check if we can extend available data by adding internal or external data sources (public or social data). The second step is linking all data throughout the organization, and make data available for machine learning analysis. In many cases data in some way is corrupted and we put effort in combining and joining all data files. The third step is analyzing the connected data and evaluate possible algorithms, based on various machine learning tools, for example:
- Decision Tree Algorithms
- Artificial Neural Network Algorithms
- Clustering Algorithms
- Instance based Algorithms
- Deep Learning Algorithms
After modelling, the fourth step is to check results with available theories, in order to compare with earlier research and common sense. Finally we need to implement changes and keep track of outcomes. Testing is neccesary to check whether learnings can provide actionable insights that improve your organization. Often A/B testing is used to assess and evaluate.
For more information CONTACT US
