in an Excel-based implementation of genetic algorithms for
optimizing binary, integer, real, and permutation valued functions that you can create
on any Excel spreadsheet. Genetic algorithms are inspired by Darwinian evolution:
over time, an initial population of answers improve and converge to
a population of optimal answers.
Case-Based Reasoning and Expert Systems
Induce-It creates case-based reasoning
expert systems from Microsoft Excel spreadsheet databases. Induce-It
searches a case database based on
similarity metrics. Case-based are
adapted from the closest matching cases, ranked by case score, and displayed
to users in a sorted list. Who Uses Induce-It?
Machine Learning and Neural Networks
is an open source(Apache 2.0 license), functional programming specification of backpropagation that yields a visual and transparent implementation within spreadsheets. This backpropagation implementation exploits array worksheet formulas, manual calculation, and has a sequential order of computation similar to the processing of a systolic array. The implementation uses no hidden macros nor user-defined functions; there are no loops, assignment statements, or links to any procedural programs written in conventional languages.
NNetSheet (currently retired) was our legacy neural network implementation (algorithms for perceptron, delta rule, generalized delta rule with back propagation and unsupervised learning (Hamming, Fuzzy Hamming, and Euclidean clustering).
RunRandom is used for quasi-random Monte Carlo simulation.
RunRandom spreadsheet formulas extend the Excel =RAND function: It
Generates quasi-random vectors.
Quasi random vectors have better convergence properties
in many Monte Carlo simulation applications.
RunPCA (currently retired) is an application -- limited by dynamic memory --
for principal components analysis.
Principal component analysis useful for
reducing the complexity of high dimensional data: a high dimensional data set can
be approximated with fewer dimensions. PCA is used in datamining
and for pre-processing input data for neural networks and regression.
KernelNet (currently retired) is used for non-parametric
multivariate kernel regression
(also known as General Regression Neural Networks). It is used for pattern
recognition and forecasting that is based on discovering the underlying probability density of
the observed data. KernelNet uses a variation of Sliced
Inverse Regression (SIR) for dimension reduction
to improve efficiency.
Purchase Student Versions
(c) 2019 Inductive Solutions, Inc. All rights