Data analysis to optimize an explicative rule for analyzing the operation of a production line

1. A method implemented by a computer, including a processor that has been programmed with instructions that cause the computer to optimize an explicative rule used for analyzing the operation of a production line, comprising:

collecting data by the computer to constitute a database of examples in the form of a table in which each line of the table is an example and each column of the table is a variable of the example;

storing the table in a computer-readable storage medium;

constructing at least one initial explicative rule relating to the table, the at least one initial explicative rule including at least one logical combination of at least one elementary conclusion, each elementary conclusion including a restriction of the domain of one of the variables;

calculating values relating to quality indicators describing a pertinence of the initial explicative rule when applied to the examples of the database;

determining at least one modified explicative rule optimizing a value of the quality indicators of the at least one initial explicative rule; and

displaying a graphical representation of the at least one modified explicative rule and corresponding quality indicators on a display, wherein the axes of the graphical representation correspond to at least two selected quality indicators.

2. The method according to claim 1 , further comprising:

selecting from among a set of said quality indicators at least one quality indicator of the at least one initial explicative rule the value of which must be optimized.

3. The method according to claim 2 , wherein calculation of the at least one modified explicative rule also minimizes the distance to the at least one initial rule.

4. The method according to claim 1 , wherein the at least one initial explicative rule further includes at least one elementary premise including a restriction of the domain of a variable.

5. The method according to claim 1 , wherein the at least one initial explicative rule includes a logical combination of elementary premises, each elementary premise including a restriction of the domain of a variable.

6. The method according to claim 1 , wherein one of the quality indicators is a rule size indicator that characterizes the number of examples of the database of examples that verify premises of the at least one initial explicative rule.

7. The method according to claim 1 , wherein one of the quality indicators is a rule volume indicator that characterizes the volume of a support subspace of the at least one initial explicative rule.

8. The method according to claim 1 , wherein one of the quality indicators is a rule purity indicator that characterizes a proportion among examples that verify premises of the at least one initial explicative rule of those that also verify its conclusions.

9. The method according to claim 1 , wherein one of the quality indicators is a rule sensitivity indicator that characterizes resistance of the at least one initial explicative rule to changes.

10. The method according to claim 1 , wherein one of the quality indicators is a robustness indicator that characterizes a nonrandom character of the at least one initial explicative rule.

11. The method according to claim 1 , further comprising:

displaying a graphical representation of the at least one modified explicative rule and corresponding quality indicators.

12. The method according to claim 1 , further comprising:

displaying a graphical representation of the result of application of the at least one modified explicative rule on the database of examples.

13. The method according to claim 1 , wherein determining the objectives includes determining thresholds of values of the quality indicators.

14. The method according to claim 1 , further comprising:

selection and invariance of a part of elementary premises of an initial rule.

15. The method according to claim 1 , further comprising:

estimating and presenting to the user the time required for optimization processing.