Center for Machine Learning and Intelligent Systems
About  Citation Policy  Donate a Data Set  Contact

Repository Web            Google
View ALL Data Sets

× Check out the beta version of the new UCI Machine Learning Repository we are currently testing! Contact us if you have any issues, questions, or concerns. Click here to try out the new site.

Function Finding Data Set
Download: Data Folder, Data Set Description

Abstract: Cases collected mostly from investigations in physical science; intention is to evaluate function-finding algorithms

Data Set Characteristics:  


Number of Instances:




Attribute Characteristics:


Number of Attributes:


Date Donated


Associated Tasks:


Missing Values?


Number of Web Hits:




Cullen Schaffer
Department of Computer Science
Rutgers University
New Brunswick, NJ 08903
schaffer '@'


Cullen Schaffer, Domain-Independent Scientific Function Finding.
PhD Thesis, Department of Computer Science, Rutgers University, 1990 (Technical Report LCSR-TR-149).

Data Set Information:

[Please note the use of Latex format here for algebraic expressions. See Leslie Lamport, Latex: A Document Preparation System, Addison-Wesley, 1986 for details.]

This database contains 352 bivariate numeric data sets collected from diverse sources and resulting, with a few exceptions, from investigations in physical science. For each data set, the collection includes:

1. Source: Bibliographic information for the source of the data.
2. Description: Identification of the variables $x$ and $y$. Except in a few clearly identified instances, the abbreviated format $y$ vs. $x$ is employed. An entry of the form

Description: Force vs. separation.

indicates that $x$ is a separation and $y$ is a force. In some cases--when the information was readily available--the description also includes the units in which the data was originally reported.

3. Reference relation: The functional relationship proposed by the reporting scientist in the original source.

4. Comments (optional): Additional information pertaining to the case.

In recording reference relations, the database often omits details of parameter values. If a scientist proposes $y=23.1x-.0014$, the reference relation may be given as just $y=k_{1}x+k_{2}$. Also, since algebraic transformations have been employed freely, the same relation might be given as $y/x=k_{2}/x+k_{1}$.

In general, data collected here is given in full as it appeared in the original source. Fractions have been converted to decimals, numbers have been freely translated to and from scientific notation and zeros have sometimes been added to decimal numbers to facilitate tabulation. Any additional deviations from verbatim transcription are noted in the Comments entry of the associated case. Note in particular that, in a few clearly identified cases, apparent typographical errors have been corrected and that, in others, data points identified by the reporting scientist as *not* conforming to the proposed relationship have been omitted.

Attribute Information:


Relevant Papers:

Cullen Schaffer, "A Proven Domain-Independent Scientific Function-Finding Algorithm," in AAAI-90.
[Web Link]

Citation Request:

Please refer to the Machine Learning Repository's citation policy

Supported By:

 In Collaboration With:

About  ||  Citation Policy  ||  Donation Policy  ||  Contact  ||  CML