Conservation dependencies
AssigneeAT&T Intellectual Property I, L.P.
InventorsLukasz Golab, Howard Karloff, Philip Korn, Divesh Srivastava, Barna Saha
Life of this patent
Estimating the term…
Abstract
Given a set of data for which a conservation law is an appropriate characterization, “hold” and/or “fail” tableaux are provided for the underlying conservation law, thereby providing a conservation dependency whereby portions of the data for which the law approximately holds or fails can be discovered and summarized in a semantically meaningful way.
Claims4
A computer-implemented method of processing a data set, wherein the data set comprises a pair of numerical sequences a ={a1, a 2, . . . ai, . . . an} and b ={b1, b 2, . . . bi. . . bn,} with ai, bi ≧0 for all integer values of i, and for which the ith pair of values, (ai, bi), is associated with the ith value, ti, of an ordered attribute t ={t1, t 2, . . . ti, . . . tn}, the method comprising:
- generating, by a processor, a cumulative total by time t of the a of the pair of numerical sequences to a time T;
- generating, by the processor, the cumulative total over the time t of the b of the pair of numerical sequences to the time T;
- defining, by the processor, a tableau as subsets of the ith pair of values (ai, bi) having confidences;
- determining, by the processor, the tableau as one of a hold tableau and a fail tableau, the hold tableau having all the confidences at least equal to a minimum value, and the fail tableau having all the confidences less than or equal to a maximum value;
- generating, in storage, an electronic plot of the cumulative total over the time t of the a of the pair of numerical sequences;
- generating, by the processor, a cumulative plot by plotting the cumulative total over the time t of the b of the pair of numerical sequences onto another plot of the cumulative total over the time t of the a of the pair of numerical sequences;
- adding, by the processor, a baseline to the cumulative plot as the cumulative total over the time t of the a of the pair of numerical sequences, where t ={t1, t2, . . . tn-1};
- determining, by the processor, an area between plots of the cumulative total over the time t of the pair of the numerical sequences a and b;
- determining, by the processor, a balance model using the area between the plots of the cumulative total over the time t of the pair of the numerical sequences a and b;
- comparing, by the processor, the cumulative total over the time t of the a of the pair of numerical sequences to the cumulative total over the time t of the b of the pair of numerical sequences;
- determining, by the processor, a conservation law in response to the comparing of the cumulative total over the time t of the a of the pair of numerical sequences to the cumulative total over the time t of the b of the pair of numerical sequences;
- determining, by the processor, a smallest hold tableau having a smallest value of all the confidences that are at least equal to the minimum value;
- and
- defining, by the processor, a conservation dependency as the smallest hold tableau associated with the a of the pair of numerical sequences and the b of the pair of numerical sequences, thus satisfying the conservation law.
A memory storing instructions that when executed cause a processor to perform operations, the operations comprising:
- generating a cumulative total by time t of a sequence a of a pair of numerical sequences, wherein the sequence a ={a1, a 2, . . . ai, . . . an}, with ai ≧ 0 for all integer values of i;
- generating a corresponding cumulative total over the time t of a sequence b, wherein the numerical sequence b ={b1, b 2, . . . bi . . . bn} with bi ≧ 0 for the all integer values of i and for which the ith pair of values, (ai, bi), is associated with the ith value time ti of an ordered attribute t ={t1, t2, . . . ti, . . . tn};
- defining a tableau as subsets of the ith pair of values (ai, bi) having confidences;
- determining the tableau as one of a hold tableau and a fail tableau, the hold tableau having all the confidences at least equal to a minimum value, and the fail tableau having all the confidences less than or equal to a maximum value;
- generating an electronic plot of the cumulative total over the time t of the numerical sequence a;
- plotting the corresponding cumulative total over the time t of the numerical sequence b onto the electronic plot of the cumulative total over the time t of the numerical sequence a;
- comparing the cumulative total over the time t of the numerical sequence a to the cumulative total over the time t of the numerical sequence b;
- determining a conservation law in response to the comparing of the cumulative total over the time t of the numerical sequence a to the cumulative total over the time t of the numerical sequence b;
- determining a smallest hold tableau having a smallest value of all the confidences that are at least equal to the minimum value;
- and
- defining a conservation dependency as the smallest hold tableau associated with the a of the pair of numerical sequences and the b of the pair of numerical sequences, thus satisfying the conservation law.
Dependencies are inferred from the claim text.