Kinaxis has been granted a patent for systems and methods that involve engineering features by receiving and fusing internal and external signals. The features are generated based on valid combinations and selected based on their predictive strength. These selected features can be used to train and select a machine learning model for forecasting purposes. GlobalData’s report on Kinaxis gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
According to GlobalData’s company profile on Kinaxis, Predictive modeling techniques was a key innovation area identified from patents. Kinaxis's grant share as of September 2023 was 20%. Grant share is based on the ratio of number of grants to total number of patents.
The patent is granted for a computer-implemented method for feature engineering and forecasting
A recently granted patent (Publication Number: US11775996B2) describes a computer-implemented method for forecasting using machine learning models. The method involves engineering one or more features and selecting a forecasting method based on a forecast request.
To engineer the features, the method involves receiving internal and external signal data, fusing the data based on metadata, generating multiple features using valid combinations from a library of transformations, and selecting features based on their predictive strength.
The forecasting method is chosen by training one or more machine learning models using the selected features. The method provides three options for choosing the model: selecting a model from the trained models, retraining a previously-selected model, or using the previously-selected model for the forecast.
The patent also describes different scenarios for selecting the machine learning model based on the forecast request. If it is the first request, the models are trained on a portion of the dataset, validated on another portion, and then retrained on the entire dataset.
If new categories of processed data have been added since the last forecast, the models are trained on an expanded dataset that includes the new data.
If an additional amount of processed data exceeds a threshold since the last forecast, the models are trained on the expanded dataset that includes the additional data.
In cases where the forecast accuracy falls below a threshold, the models are trained on an expanded dataset that includes incoming processed historical product data.
If the time interval between the most recent forecast and the request exceeds a threshold, the previously-selected model is retrained on an expanded dataset that includes new processed data collected during the interval.
The patent also mentions training different configurations of the machine learning models, such as varying the number of layers in a neural network.
Overall, this patent describes a computer-implemented method for forecasting that involves engineering features, selecting machine learning models, and adapting the models based on different scenarios and forecast requests.
To know more about GlobalData’s detailed insights on Kinaxis, buy the report here.
Data Insights
From
The gold standard of business intelligence.
Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.
GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.
GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.
