The SAS Data Science Blog
Advanced analytics from SAS data scientists
An embedding model is a way to reduce the dimensionality of input data, such as images. Consider this to be a type of data preparation applied to image analysis. When an embedding model is used, input images are converted into low-dimensional vectors that can be more easily used by other computer vision tasks. The key to good embedding is to train the model so that similar images are converted to similar vectors.
I spoke with Antonie Berkel and Dr. Joost Huiskens about a systems analysis project for the design of a clinical decision support system (CDSS) which integrates computer vision to help radiologists monitor the progression of tumors.
In the preceding two posts, we looked at issues around interpretability of modern black-box machine-learning models and introduced SAS® Model Studio within SAS® Visual Data Mining and Machine Learning. Now we turn our attention to programmatic interpretability.
In the second of a three-part series of posts, SAS' Funda Gunes and her colleague Ricky Tharrington summarize model-agnostic model interpretability in SAS Viya.
A monotonic relationship exists when a model’s output increases or stays constant in step with an increase in your model’s inputs. Relationships can be monotonically increasing or decreasing with the distinction based on which direction the input and output travel. A common example is in credit risk where you would expect someone’s risk score to increase with the amount of debt they have relative to their income.
We will use prescriptive analytics and optimization to select a stock portfolio that maximizes returns while taking risk tolerance into account.