Predictive analytics in the Trendskout AI Platform

Prediction through machine learning or deep learning can be done in a number of different ways, depending on the underlying algorithm that is used. As the name suggests, predictive models are designed to predict unknown values, properties or events. They often rely on algorithms designed for classification, clustering, pattern recognition and image recognition. The following types are natively included in Trendskout:

Application Examples

Powerful Cloud-AI Platform

In addition to specific modules for manufacturers, the Trendskout AutoML platform also offers many possibilities for IT teams to quickly set up AI applications, Powered by our unique double AI low and easy to use UI that fully automates data transformations, algorithm selection and hypertuning.

How do organizations benefit from AI and Machine Learning?

Gain access to our Resource Center with interesting business cases and how-to's.

Connect

1

Analysis

2

Prediction

Predictie is one of the available Trendskout AI Flow analysis-functions.

Automate

3

How does this work technically?

Forecasting

A common application is predicting a continuous series of values. With forecasting, the data is continuous, i.e. the values ​​are in a constant stream. This is in contrast to the prediction of labels or types where the values ​​are discrete, so they do not continuously follow each other. Temperature is an example of a continuous series of values, a temperature can be measured at any time. So predicting what the temperature will be tomorrow, for example, is forecasting. Predicting whether there will be a thunder storm tomorrow is labeling, after all, this does not happen every day. There are different types of algorithms that are used, such as SMOreg. Different data operations are required on the supplied data, and other algorithms may be better suited based on the outcome. The data operations and the evaluation of the algorithm are determined by Trendskout, so no intervention is required from the user. Here, just as with other applications such as classification, a training step is always used in which the Forecasting algorithm and its hyper-tuning are controlled by evaluation based on the accuracy of the prediction on previously processed test data, part of the training data.

Anomaly prediction

Anomaly prediction is the detection and prediction of exceptional events, anomalies. Compared to label prediction, and certainly forecasting, anomalies are events that occur even less frequently, such as the occurrence of a hurricane. Because these events occur so rarely, there is often not much information in the training data on which the algorithm can rely to detect relationships when these anomalies do or do not occur. In other words, there is an imbalance in the training data. As a result, other types of algorithms, and data processing, are required. In the hypertuning process of anomaly prediction, techniques such as SMOTE, undersampling or oversampling are frequently used. The entire hypertuning process is automated just like with all other AI and Deep Learning applications within Trendskout.

Prediction of labels or types

Although the ultimate goals of every prediction application are very similar, predicting values ​​or events, the technique for achieving this is very different. There are different variables such as the choice of the algorithm or type of neural network, the parameters of this algorithm and the required data operations. These three variables together form an infinite number of possibilities, each with a different predictive model as a result. Via intelligent Auto ML, automated data processing and a powerful intuitive user interface, Trendskout will independently select the most optimal predictive model.

Predictive Analytics + Trendskout

The above examples are just a sample of the number of prediction applications supported by Trendskout.

Although the ultimate goals of every prediction application are very similar, predicting values ​​or events, the technique for achieving this is very different. There are different variables such as the choice of the algorithm or type of neural network, the parameters of this algorithm and the required data operations. These three variables together form an infinite number of possibilities, each with a different predictive model as a result. Via intelligent Auto ML, automated data processing and a powerful intuitive user interface, Trendskout will independently select the most optimal predictive model.

Ready to discover all features during a live demo?

Get in touch and we will be happy to show you the direct business value of artificial intelligence for your organisation.