tag:blogger.com,1999:blog-5737874959005852552.post8953516432624621587..comments2023-10-18T16:31:41.080-05:00Comments on MVPA Meanderings: linear SVM weight behaviorJo Etzelhttp://www.blogger.com/profile/04277620767760987432noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-5737874959005852552.post-28118340471738287812012-07-05T09:55:29.954-05:002012-07-05T09:55:29.954-05:00I believe the best way to interpret weights to the...I believe the best way to interpret weights to the problem domain is via the so called "Kohonen networks". They were originally proposed for unsupervised learning, but supervised versions exist now.The AI bullyhttps://www.blogger.com/profile/04594432604835317193noreply@blogger.com