Optimizing tree pruning for decision trees
A system and method of optimizing tree pruning for a decision tree may include splitting, a first dataset into a training dataset and a testing dataset, growing the training dataset into a first decision tree, sampling the training dataset by creating a plurality of sampling datasets from the training dataset, pruning the first decision tree using the plurality of sampling datasets, creating a plurality of models, at least one for each of the plurality of sampling datasets, and verifying the accuracy of each of the plurality of models, using the testing dataset.
1. A method of optimizing tree pruning for a decision tree, the method comprising:
splitting, by one or more processors of a computer system, a first dataset into a training dataset and a testing dataset;
growing, by the one or more processors of the computer system, the training dataset into a first decision tree;
sampling, by the one or more processors of the computer system during the growing, the training dataset by creating a plurality of sampling datasets from the training dataset;
pruning, by the one or more processors of the computer system, the first decision tree using the plurality of sampling datasets;
creating, by the one or more processors of the computer system, a plurality of models, at least one for each of the plurality of sampling datasets; and
verifying, by one or more processors of a computer system, the accuracy of each of the plurality of models, using the testing dataset.
2. The method of claim 1 , wherein each of the plurality of datasets comprises one machine cache.
3. The method of claim 1 , further comprising determining, by the one or more processors of the computer system, a most accurate model from the plurality of models.
4. The method of claim 1 , wherein the pruning is completed on a distribution cluster such that the pruning of each of the plurality of sampling datasets occurs on a different machine, and wherein the growing is completed on the distribution cluster such that the growing the training dataset into the first decision tree includes growing each of the plurality of sampling datasets on a different machine.
5. The method of claim 1 , wherein the sampling further includes important sampling and sampling with replacement.
6. The method of claim 1 , wherein the verifying the accuracy of each of the plurality of models begins as soon as a first of the plurality of models are created, wherein verifying the accuracy of each of the plurality of models occurs in the order each of the plurality of models is created.
7. The method of claim 1 , wherein the sampling occurs on a first machine of a distribution cluster, and wherein the pruning of each of the plurality of sampling datasets occurs on one or more machines that are not the first machine.
8. A computer system, comprising:
one or more processors;
one or more memory devices coupled to the one or more processors; and
one or more computer readable storage devices coupled to the one or more processors, wherein the one or more storage devices contain program code executable by the one or more processors via the one or more memory devices to implement a method of optimizing tree pruning for a decision tree, the method comprising:
splitting, by the one or more processors of the computer system, a first dataset into a training dataset and a testing dataset;
growing, by the one or more processors of the computer system, the training dataset into a first decision tree;
sampling, by the one or more processors of the computer system during the growing, the training dataset by creating a plurality of sampling datasets from the training dataset;
pruning, by the one or more processors of the computer system, the first decision tree using the plurality of sampling datasets;
creating, by the one or more processors of the computer system, a plurality of models, at least one for each of the plurality of sampling datasets; and
verifying, by one or more processors of a computer system, the accuracy of each of the plurality of models, using the testing dataset.
9. The computer system of claim 8 , wherein each of the plurality of datasets comprises one machine cache.
10. The computer system of claim 8 , the method further comprising determining, by the one or more processors of the computer system, a most accurate model from the plurality of models.
11. The computer system of claim 8 , wherein the pruning is completed on a distribution cluster such that the pruning of each of the plurality of sampling datasets occurs on a different machine, and wherein the growing is completed on the distribution cluster such that the growing the training dataset into the first decision tree includes growing each of the plurality of sampling datasets on a different machine.
12. The computer system of claim 8 , wherein the sampling further includes important sampling and sampling with replacement.
13. The computer system of claim 8 , wherein the verifying the accuracy of each of the plurality of models begins as soon as a first of the plurality of models are created, wherein verifying the accuracy of each of the plurality of models occurs in the order each of the plurality of models is created.
14. The computer system of claim 8 , wherein the sampling occurs on a first machine of a distribution cluster, and wherein the pruning of each of the plurality of sampling datasets occurs on one or more machines that are not the first machine.
15. A computer program product, comprising a computer readable hardware storage device storing a computer readable program code, the computer readable program code comprising an algorithm that when executed by one or more processors of a computer system implements a method of optimizing tree pruning for a decision tree, the method comprising:
splitting, by the one or more processors of the computer system, a first dataset into a training dataset and a testing dataset,
growing, by the one or more processors of the computer system, the training dataset into a first decision tree;
sampling, by the one or more processors of the computer system during the growing, the training dataset by creating a plurality of sampling datasets from the training dataset;
pruning, by the one or more processors of the computer system, the first decision tree using the plurality of sampling datasets;
creating, by the one or more processors of the computer system, a plurality of models, at least one for each of the plurality of sampling datasets; and
verifying, by one or more processors of a computer system, the accuracy of each of the plurality of models, using the testing dataset.
16. The computer program product claim 15 , wherein each of the plurality of datasets comprises one machine cache.
17. The computer program product of claim 15 , the method further comprising determining, by the one or more processors of the computer system, a most accurate model from the plurality of models.
18. The computer program product of claim 15 , wherein the pruning is completed on a distribution cluster such that the pruning of each of the plurality of sampling datasets occurs on a different machine, and wherein the growing is completed on the distribution cluster such that the growing the training dataset into the first decision tree includes growing each of the plurality of sampling datasets on a different machine.
19. The computer program product of claim 15 , wherein the sampling further includes important sampling and sampling with replacement.
20. The computer program product of claim 15 , wherein the verifying the accuracy of each of the plurality of models begins as soon as a first of the plurality of models are created, wherein verifying the accuracy of each of the plurality of models occurs in the order each of the plurality of models is created.