System and method for ranking objects by likelihood of possessing a property
First Claim
Patent Images
1. A method for ranking objects by the likelihood of possessing a property, comprising:
- a) storing electronic objects in at least one database in memory of a computer system;
b) obtaining a training set of objects with a sub-population known to possess property (Y) and with a sub-population known to possess non-property (Ybar) which is indicative of not possessing property (Y);
c) determining a likelihood that objects (q) within a population (Q) of the stored objects relate to at least one of said known property and known non-property;
d) ranking said objects and for similarity to said training set of objects;
e) generating a report comprising the ranking of the objects for display, transmission, or storage in memory.
2 Assignments
0 Petitions
Accused Products
Abstract
A system and method for ranking objects by a likelihood of possessing a property is disclosed. The system can be used, for example, to assist in the determination of system events, such as, e.g., computer system failures, based upon an analysis of customer service reports or the like.
14 Citations
25 Claims
-
1. A method for ranking objects by the likelihood of possessing a property, comprising:
-
a) storing electronic objects in at least one database in memory of a computer system; b) obtaining a training set of objects with a sub-population known to possess property (Y) and with a sub-population known to possess non-property (Ybar) which is indicative of not possessing property (Y); c) determining a likelihood that objects (q) within a population (Q) of the stored objects relate to at least one of said known property and known non-property; d) ranking said objects and for similarity to said training set of objects; e) generating a report comprising the ranking of the objects for display, transmission, or storage in memory. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A computer system for facilitating the determination of the likelihood that an object possesses a property, comprising:
-
means for storing electronic objects in at least one database in memory of the computer system; means for obtaining a training set of objects related at least one of known property (Y) and known non-property (Ybar), wherein relation to the known non-property is indicative of not possessing the known property; means for determining a likelihood that objects (q) within a population (Q) relate to at least one of said known property and known non-property; means for isolating features (f) possessed by objects in said training set of objects, wherein the features are surrogates for the known property; means for identifying the significance (s(f)) of said features (f) in distinguishing possession or non-possession of the known property; means for selecting a subset (Fs) of a union (Fp) of said features (f) based upon said significance; means for computing a similarity indicator (ss(q)) for each object (q) to be ranked using the presence or absence of each feature (f) in said subset (Fs); means for ranking said objects (q) for similarity to said training set of objects based on the determined likelihood and at least partially on the similarity indicator (ss(q)) values; and means for generating a report comprising the ranking of the object that is stored in the memory, transmitted over a communications network, or displayed on a display device. - View Dependent Claims (17, 18, 19, 20, 21)
-
-
22. A computer system for facilitating the determination of system events, comprising:
-
a) at least one input unit for inputting information related to cases; b) at least one database for storing electronic documents based on information inputted via said at least one input unit, including a training set of documents related to at least one of known events (Y) and known non-events (Ybar); c) a processor coupled to said at least one database and to data storage; d) wherein said data storage is programmed to cause said processor to determine a likelihood that objects (q) within a population (Q) relate to at least one of said known events and known non-events; e) wherein said data storage is programmed to cause said processor to rank said objects (q) for similarity to said training set of objects to indicate likelihood of the known events being present in the ranked objects, the ranking being based also on the similarity of the ranked objects to the training set of documents related to the known non-events; and f) wherein said processor storesstoring a report comprising the rank of said objects. - View Dependent Claims (23, 24, 25)
-
Specification