Methods and systems for enabling analysis of communication content while preserving confidentiality

Abstract

Disclosed are methods and systems for enabling analysis of communication content while preserving confidentiality. In one embodiment, communication content is processed to increase the similarity of superficially dissimilar instances of communication content and/or to increase the distinctiveness of superficially similar instances of communications content. In this embodiment at least part of the processed communication content is hashed to obscure the actual communication content. In one embodiment, social network analysis is performed on the communication content after hashing, and visualization of the social network analysis includes thread graphs and/or circular graphs.

1 - 25 . (canceled) 
   
   
       26 . A system for enabling analysis of communication content while preserving confidentiality, comprising:
 means for capturing communication content including instances of communication content that can be rendered into text;   means for processing said captured communication content into natural language tokens to adjust a level of similarity between separate instances of communication content, wherein each natural language token represents a root stem; and   means for hashing at least part of said processed communication content to obscure the actual communication content and to produce hashed tokens corresponding to each natural language token, wherein the level of similarity between separate instances of communication content is adjusted to improve hashing results.   
   
   
       27 . The system of  claim 26 , wherein said means for processing said captured communication content executes at least one of: encrypting, decrypting, deleting spam, correcting typographical errors, identifying communications related to the same social unit, identifying idiomatic expressions and diagramming sentence structure, and dropping stop words. 
   
   
       28 . The system of  claim 26 , wherein said means for hashing obscures at least one of the following associated with the communication content: an author of the communication, a recipient of the communication, a topic of the communication, and a body of the communication. 
   
   
       29 . The system of  claim 26 , further comprising:
 means for analyzing said communication content, subsequent to said hashing, wherein the analysis includes determining similar keywords using information retrieval techniques.   
   
   
       30 . The system of  claim 29 , wherein said means for analyzing analyzes the hashed tokens of the processed communication content for similarity with another captured communication. 
   
   
       31 . The system of  claim 29 , further comprising:
 means for outputting at least one selected from a group comprising: message analysis, usage analysis, diffusion analysis, and network analysis.   
   
   
       32 . The system of  claim 31 , wherein said means for outputting usage analysis employs an interface providing interaction visualization among at least two social units across a time period. 
   
   
       33 . The system of  claim 31 , wherein said means for outputting network analysis employs a graphical interface having a circular layout where information on a social network is provided by at least one parameter selected from a group comprising: measure of radii of nodes representing social units in said network, measure of angles of nodes representing social units in said network, color of nodes representing social units in said network, color of links connecting nodes representing social units in said network, shading of nodes representing social units in said network, shading of links connecting nodes representing social units in said network, line-type of links connecting nodes representing social units in said network, and shading of a range of angles of nodes representing social units in said network. 
   
   
       34 . The system of  claim 31 , wherein said means for outputting network analysis includes a filter for dynamically changing the size of a social network that is being visualized. 
   
   
       35 . The system of  claim 31 , wherein said means for outputting network analysis includes a filter for dynamically changing the threshold of connectivity of a social network which is being visualized. 
   
   
       36 . The system of  claim 26 , wherein said means for hashing produces a plurality of hashed tokens for each entire communication included in said captured communication content. 
   
   
       37 . The system of  claim 26 , wherein said means for processing said captured communication content increases a level of similarity of separate instances of communication content which superficially appear to be dissimilar, and wherein said means for hashing produces similar hashed tokens for similar instances of processed communication content. 
   
   
       38 . The system of  claim 26 , wherein said means for processing said captured communication content reduces a level of similarity of separate instances of communication content which superficially appear to be similar, and wherein said means for hashing produces dissimilar hashed tokens for dissimilar instances of processed communication content. 
   
   
       39 . A method of enabling analysis of similarity of instances of communication content while preserving confidentiality, comprising:
 capturing communication content including instances of communication content that can be rendered into text;   processing said captured communication content into natural language tokens to adjust a level of similarity between separate instances of communication content, wherein each natural language token represents a root stem; and   hashing at least part of said processed communication content to obscure the actual communication content and to produce hashed tokens corresponding to each natural language token, wherein the level of similarity between separate instances of communication content is adjusted to improve hashing results.   
   
   
       40 . The method of  claim 39 , wherein said processing includes at least one from a group comprising: correcting typographical errors, identifying separate instances of communication content related to a same social unit even though said separate instances of communication content appear to be related to different social units, identifying idiomatic expressions and diagramming sentence structure, dropping stop words, and applying morphological techniques to reduce the dissimilarity of similar words and expressions. 
   
   
       41 . The method of  claim 40 , wherein said applying morphological techniques includes at least one selected from a group comprising: dropping prefixes, dropping suffixes, root stemming nouns, reducing irregular verbs to a single base, and eliminating past, present, and future tenses. 
   
   
       42 . The method of  claim 39 , wherein said obscuring the actual content of said communication content includes obscuring at least one selected from a group comprising: an author of said communication content, a recipient of said communication content, a topic of said communication content, and at least part of a body of said communication content. 
   
   
       43 . The method of  claim 39 , wherein part of said processed communication content which does not compromise personal privacy is retained unhashed in order to facilitate any later analysis of said communication content. 
   
   
       44 . The method of  claim 39 , wherein said hashing includes rearranging a word order of said communication content. 
   
   
       45 . The method of  claim 39 , further comprising: analyzing said communication content, subsequent to said hashing. 
   
   
       46 . The method of  claim 45 , wherein said analyzing includes analyzing data related to social networks including at least one selected from a group comprising: degree of collaboration, level of information proximity, level of knowledge exchange, differences in behavior by status, differences in effectiveness correlated with differences in use of communications technology, network tie strength, how information flows affect social unit output, connections between information and productivity, and any differences in productivity based on how social units use information. 
   
   
       47 . The method of  claim 45 , wherein said analyzing includes comparing and classifying said hashed tokens using at least one method selected from a group comprising: statistics, linguistic structure analysis, information distance metrics, syntactic, and semantic cues analysis. 
   
   
       48 . The method of  claim 47 , wherein said captured communication content is from email communication and wherein said analyzing includes a cosine metric, wherein said cosine metric includes assigned weights which take into account a structure of email communication. 
   
   
       49 . A method of analyzing the similarity of communications while preserving the confidentiality of the communications, comprising:
 capturing at least two entire communications;   processing the at least two entire communications into natural language tokens to improve the similarity of any similar content within the at least two entire communications and to reduce the similarity of any dissimilar content within the at least two entire communications, wherein each natural language token represents a root stem;   encrypting the at least two processed communications to generate tokens which obscures the actual content and are similar in nature for similar content, wherein each generated token corresponds to a natural language token; and   comparing the tokens to identify similar content within the at least two processed communications without determining the actual content of the least two processed communications.   
   
   
       50 . The method of  claim 49 , wherein the encrypting is adjusted to provide a first specific level of actual content privacy and a second specified level of comparison similarity between the at least two processed communications. 
   
   
       51 . The method of  claim 50 , wherein the encrypting is increased to improve actual content privacy and reduce comparison similarity.