Description
ndCurveMaster is a machine learning application designed to find the best complex nonlinear equations to describe empirical data by using a multiple nonlinear regression method, and:
• can load a data set consisting of measured values and automatically discovers regression equation set,
• can auto fit an unlimited number of input variables and their combinations, for example
• utilizes an automated machine learning method for the discovery of equations,
• offers a multicollinearity detection option through the use of a variance inflation factor (VIF) and Pearson Correlation Matrix to improve model quality,
• offers an overfitting detection option,
• uses heuristic techniques to find nonlinear equations,
• quickly generates high quality results and saves time.
You may try the product free for 9 days.
-
fully automated process curves to fit your data set
-
automated discovery of nonlinear equations
-
auto data fitting of an unlimited number of fitted variables and their combinations
-
any order polynomial fitting for any number of variables, a set of templates for polynomial functions
-
heuristic techniques for data fitting by the use of random iterated or full random search algorithms
-
full statistical analysis of every equation: analysis of variance (ANOVA), regression and collinearity
-
multicollinearity detection through the use of a Variance inflation factor (VIF) and Pearson Correlation Matrix
-
search for models with a VIF limit value
-
weight data for individual data points in any column
-
detect and prevent overfitting
-
models with or without an intercept term
-
free expansion or reduction of the model by the user
-
import data from CSV, TXT, XLSX, XLS spreadsheets
-
result history and ranking
-
up to 380 basic functions such as power, exponential, logarithmic and trigonometric functions grouped in 5 collections; user-created custom function collection
-
copy, save or load results
-
high-quality plots: 2D Fitted line, regression, residual plots and histogram in graph form
-
set any significant level alpha value
-
optimized to manage large amounts of data
-
ability to lock the predictor variable option to not modify a preselected predictor variable while searching