Electronic discovery using predictive filtering
Electronic discovery using predictive filtering is disclosed herein. An example method includes providing a filtering interface that includes a field value input, a predicted values selector, and a predictor type selector; receiving at least a pivot selected from the field value input and a predicted value from the predicted values selector; searching a plurality of documents based on the pivot and the predicted value selected for any of predictive phrases or predictive concepts; calculating a predictive value for each of the predictive phrases or predictive concepts; and generating a graphical user interface that includes the predictive phrases or predictive concepts in conjunction with their respective predictive value.
1. A method, comprising:
receiving a selection of a pivot, the pivot being a field value or a set of field values for a plurality of documents;
generating a prediction of responsive phrases, responsive concepts, or other meta-data, the prediction based at least in part of evaluating a plurality of documents based on the pivot;
calculating a predictive value for each of the predicted responsive phrases or predicted responsive concepts, the predictive value being indicative of a likelihood that the predicted responsive phrases or predicted responsive concepts are to be associated with documents of the plurality of documents that are tagged with the pivot;
generating a graphical user interface that comprises automatically generated filter criteria based on the predictive value for each of the predicted responsive phrases, predicted responsive concepts, or other predictive meta-data;
receiving a selection of at least one of the automatically generated filter criteria from the graphical user interface;
building and applying a filter based on the selection; and
displaying within the graphical user interface, documents from the plurality of documents that were selected using the filter.
2. The method according to claim 1 , wherein calculating a predictive value for each of the predicted responsive phrases or predicted responsive concepts comprises using any of chi-squared statistic or pointwise mutual information.
3. The method according to claim 1 , wherein the pivot is updated with the selection of the at least one automatically generated filter criteria from the graphical user interface and the filter is rebuilt, the rebuilt filter identifying additional documents found in an updated search using the rebuilt filter.
4. The method according to claim 1 , wherein the plurality of documents have been tagged with field values.
5. The method according to claim 1 , wherein building the filter comprises generating and displaying a filtering interface that comprises a field value input, a predicted values selector based on the field value input, and a predictor type selector.
6. The method according to claim 5 , wherein the filtering interface further comprises a preview panel that comprises a summary comprising the predictive value for each of the predicted responsive phrases or predicted responsive concepts identified.
7. The method according to claim 6 , wherein the predicted responsive phrases or predicted responsive concepts are ranked within the graphical user interface according to their respective predictive value.
8. The method according to claim 7 , wherein each of the predicted responsive phrases or predicted responsive concepts are associated with a frequency count value that indicates how many times the predicted responsive phrases or predicted responsive concepts appear in the plurality of documents.
9. The method according to claim 8 , wherein the frequency count value is displayed next to the predictive value for each of the predicted responsive phrases or predicted responsive concepts.
10. A system, comprising:
a processor; and
a memory for storing executable instructions, the processor executing the instructions to:
receive a selection of a pivot, the pivot being a field value for a plurality of documents;
generate at least one predictive phrase or at least one predictive concept based on the selected pivot, for the plurality of documents;
calculate a predictive value for each of the predictive phrases or predictive concepts, the predictive value indicating a likelihood that the predictive phrase or predictive concept is associated with documents of the plurality of documents that are tagged with the pivot; and
generate a graphical user interface that comprises the predictive phrases or predictive concepts in conjunction with their respective predictive value.
11. The system according to claim 10 , wherein the processor further executes the instructions to receive a selection of at least one of the predictive phrases or predictive concepts from the graphical user interface.
12. The system according to claim 11 , wherein the processor further executes the instructions to build and apply a filter based on the selection.
13. The system according to claim 12 , wherein the processor further executes the instructions to display, within the graphical user interface, documents from the plurality of documents that were selected using the filter.
14. The system according to claim 13 , wherein the processor further executes the instructions to update the pivot with the selection and the filter is rebuilt, the rebuilt filter identifying additional documents found in an updated search using the rebuilt filter.
15. The system according to claim 14 , wherein the processor further executes the instructions to generate and display a filtering interface that comprises a field value input, a predicted values selector based on the field value input, and a predictor type selector.
16. The system according to claim 15 , wherein the filtering interface further comprises a preview panel that comprises a summary comprising the predictive value for each of the generated predictive phrases or predictive concepts.
17. The system according to claim 16 , wherein the predictive phrases or predictive concepts are ranked by the system within the graphical user interface according to their respective predictive value.
18. The system according to claim 17 , wherein each of the predictive phrases or predictive concepts are associated with a frequency count value that indicates how many times the predictive phrases or predictive concepts appear in the plurality of documents, and the frequency count value is displayed on the graphical user interface next to the predictive value for each of the predictive phrases or predictive concepts.
19. A method, comprising:
providing a filtering interface that comprises a field value input, a predicted values selector, and a predictor type selector for a plurality of documents;
receiving a selection of at least one pivot from the field value input and a selection of at least one predicted value from the predicted values selector;
searching the plurality of documents based on the selected pivot and the selected predicted value for at least one generated predictive phrase or at least one generated predictive concept;
calculating a predictive value for each of the generated predictive phrases or predictive concepts, the predictive value indicating a likelihood that the predictive phrase or predictive concept is associated with documents of the plurality of documents that are tagged with the selected pivot; and
generating a graphical user interface that comprises the predictive phrases or predictive concepts in conjunction with their respective predictive value.
20. The method according to claim 19 , wherein the filtering interface further comprises a preview panel that comprises a summary comprising the predictive value for each of the generated predictive phrases or predictive concepts.