Katerina Henclova·Oct 14Benchmarking Alternative Feature Selection Algorithms (Part 3: Results on Synthetic Problems)Six algorithms search for three planted answers, needle-in-a-haystack style.
InDATAMOLEbyKaterina Henclova·Oct 11Benchmarking Alternative Feature Selection Algorithms (Part 2: The Algorithms)The phenomenon of a dataset supporting several valid solutions keeps being rediscovered and renamed, e.g. the Rashomon effect, predictive…A response icon1A response icon1
InDATAMOLEbyKaterina Henclova·Sep 2Benchmarking Alternative Feature Selection Algorithms (Part 1: Design)How do you benchmark an algorithm whose job is to return several valid answers? The obvious way to score it turns out to be wrong.
InDATAMOLEbyKaterina Henclova·Aug 11What Is Predictive Multiplicity in Your Data And How GEMSS Deals With ItWhen your data has more than one story to tell, an ordinary feature selector will pick one and never mention the rest. GEMSS hands you the menu of several valid sparse solutions in a single fit.
InDATAMOLEbyKaterina Henclova·Jul 14Multivariate SSA: Meet the PCA for Time SeriesTL;DR Multivariate Singular Spectrum Analysis (mSSA) is a rigorous and highly effective method for time series preprocessing. You can think…A response icon1A response icon1
InDATAMOLEbyKaterina Henclova·Jul 8No Labels? No Problem: Unlocking IoT Time Series with Weakly Supervised LearningIf you work in machine learning, you know the data bottleneck all too well: we can get mountains of raw data, but getting the high-quality…
InDATAMOLEbyKaterina Henclova·May 12The assisted discovery of GEMS(S)Your feature selector picks one explanation and never mentions the alternatives. When the goal is understanding a mechanism, that choice…
