WO2013119762A1 - Outils et procédés de détermination d'index de relations sémantiques - Google Patents

Outils et procédés de détermination d'index de relations sémantiques Download PDF

Info

Publication number
WO2013119762A1
WO2013119762A1 PCT/US2013/025070 US2013025070W WO2013119762A1 WO 2013119762 A1 WO2013119762 A1 WO 2013119762A1 US 2013025070 W US2013025070 W US 2013025070W WO 2013119762 A1 WO2013119762 A1 WO 2013119762A1
Authority
WO
WIPO (PCT)
Prior art keywords
relationships
differences
baselines
descriptor
metric
Prior art date
Application number
PCT/US2013/025070
Other languages
English (en)
Inventor
Adam TREISER
Original Assignee
Treiser Adam
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US13/461,670 external-priority patent/US8341101B1/en
Priority claimed from US13/492,528 external-priority patent/US8943004B2/en
Priority claimed from US13/621,717 external-priority patent/US8478702B1/en
Priority claimed from US13/674,692 external-priority patent/US20130204823A1/en
Application filed by Treiser Adam filed Critical Treiser Adam
Publication of WO2013119762A1 publication Critical patent/WO2013119762A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising

Definitions

  • the present invention relates to characteristic-based profiling systems and, more particularly, to combining multiple points of data regarding individuals through the use of characteristics in order to determine the relationship between the individuals and a user-defined criteria.
  • Customer profiling systems are known in the art.
  • Traditional systems include consumer rewards cards, credit card purchase information, demographic profiling, behavioral profiling, and customer surveying.
  • Some businesses supplement these traditional systems with website and social media analytic tools that profile the business's fans and followers according to factors such as “likes,” “click-through rates,” and search engine queries, among others.
  • these systems attempt to determine products, promotions, and advertisements that are most likely to appeal to a specific customer or broad customer segment. This information helps
  • customer information is often collected with respect to a single business metric and may never be used to glean insights about other metrics that may be helpful to the company. This is particularly true for businesses that are growing and those that have multiple departments. Growing businesses must usually adjust or supplement its performance metrics to reflect new goals, strategies, and business operations. As a result, these businesses must understand how its customers relate to the new set of business metrics rather than, or in addition to, the ones for which the data was originally collected. Similarly, businesses with multiple departments frequently gather customer information for purposes of a department- specific metric, but fail to use that information across other departments or globally within the organization. For example, a business may have a marketing department and risk management department.
  • Sentiment analysis methods are capable of identifying the sentiment expressed in a single statement, or the overall sentiment expressed by multiple statements; however, these methods typically cannot determine whether this sentiment accurately reflects the individual's (or group's) true feelings, or simply reflects the mannerisms and style of that individual or group. For example, a certain individual may generally express strong sentiments as part of their personal style, even when they do not feel strongly about the topic they are discussing. Sentiment analysis methods may incorrectly determine that this individual has a strong sentiment towards a given topic, based on the sentiment normally expressed by that individual. Thus, there is a need for a more accurate sentiment analysis method that takes into account the baseline sentiment expressed by each individual or group, to more accurately determine that individual's or group's sentiment towards a given topic.
  • sentiment analysis methods are unable to determine which individuals matter the most to an organization. For example, a business may wish to identify which customers have the most impact on a marketing campaign. Likewise, an organization may wish to identify individuals that have absolutely no significance to it whatsoever, and should therefore be ignored. These sentiment analysis methods may be able to identify which individuals have positive or negative sentiment about the campaign. However, these methods will not be able to determine which of those individuals are also the most important to that business. For example, an organization may wish to identify individuals who make statements that influence others. A positive or negative review by one individual (for example, a technology critic) may be taken more seriously, and influence more people, than a similar review by another individual.
  • a technology critic for example, a technology critic
  • the sentiment expressed by an individual may be a reliable predictor of larger trends, even if that individual's statement is not influential.
  • certain individuals may have insights that are later expressed by large groups, even if the group is unaware of that insight.
  • an organization may wish to identify these insightful individuals, in order to determine trends and attitudes as early as possible.
  • an organization may wish to know how a positive or negative sentiment expressed by an individual corresponds to certain behavior by that same individual. For example, whether an individual who makes a negative statement about a service later discontinues that service. This may allow the organization to determine which dissatisfied customers it should offer the largest incentives to continue their service.
  • a business might also wish to know which individuals that make positive statements about a product later purchase that product. This may allow a business to identify which potential customers are most likely to make a purchase, in order to target those customers with incentives.
  • a business may also want to identify people with similar sentimental patterns and characteristic based profiles so as to better understand how to address, prepare for, and anticipate various situations, changes in situations, relationships, customer feelings, reactions, or sentiments.
  • the organization may also wish to track the change in an individual sentiment based on certain actions by the organization. For example, whether an incentive was effective in retaining a customer because it actually changed the customer's sentiment, or whether the customer has retained service because of the incentive, but remains dissatisfied. Or, whether a certain marketing campaign affects individuals' sentiment towards a product.
  • a computer-readable non-transitory storage medium having instructions which, when executed on a processor, perform a method for identifying relationships between individuals, metrics, and sub-metrics, using characteristics.
  • FIG. 1 is a block diagram of an exemplary embodiment of a characteristic-based server
  • FIG. 2 is a flowchart depicting one process for determining a relationship score for an individual relative to a metric
  • Fig. 3 is a block diagram depicting an example of relationships between characteristics, metrics, data items, and individuals;
  • Fig. 4 is a block diagram depicting an example of relationships between individuals and characteristics
  • Fig. 5 is a block diagram depicting an example score for an individual related to a metric
  • Fig. 6 is a block diagram depicting an example group of
  • Fig. 7 is a block diagram depicting an example of relationships between groups and characteristics
  • Fig. 8 is a block diagram depicting an exemplary score for a group
  • Fig. 9 is a block diagram depicting relationships used to determine sub-metric scores
  • Fig. 10 is a block diagram, depicting relationships used to determine scores for a metric
  • Fig. 11 is a block diagram, depicting relationships between groups, characteristics, sub-metrics, and a metric
  • Fig. 12 is a block diagram depicting a sample user screen displaying an individual and their score
  • Fig. 13 is a block diagram depicting a sample detail screen for an individual
  • FIG. 14 is a block diagram depicting a sample communication screen for an individual
  • Fig. 15 is a block diagram depicting an example of a general notification screen.
  • Fig. 16 is a block diagram depicting an example of a specific notification screen for an individual.
  • Fig. 17 is a flowchart depicting one process for determining baselines for relationships.
  • Fig. 18 is a block diagram depicting an example of a first set of relationships between information atoms.
  • Fig. 19 is a block diagram depicting an example of a second set of relationships between information atoms.
  • Fig. 20 is a block diagram depicting an example of differences between baselines and relationships.
  • Fig. 21 is a block diagram depicting an example of identified relations between a difference and other information atoms.
  • the characteristic-based system may define a number of characteristics.
  • the term characteristic broadly refers to any attribute, trait, value, or other factor associated, whether objectively or subjectively, with an individual or group of individuals. The detailed description below provides further examples of such characteristics.
  • the characteristic- based system may use a suitable relationship-determining module (i.e., a software component, a hardware component, or a combination of a software component and a hardware component) comprising relationship-determining algorithms known in the art to determine the relationship between the information and the characteristics.
  • This relationship may be described using both a magnitude and a direction. Further, the description may be represented by a numerical value, textual identifier, graphical icon, color, opacity, or any other suitable method of representing the relationship.
  • the magnitude may represent how strongly the information is related to the characteristics, including the lack of any relationship at all.
  • the relationship may also be identified as positive, negative, or neutral.
  • the term "positive” broadly refers to relationships where the existence of, or a change in, one member of the relationship corresponds to a similar existence of, or a similar change in, the other members.
  • negative broadly refers to relationships where the existence of, or a change in one member of the relationship corresponds to a lack of the existence of, or an inverse change in, the other members.
  • the system may use the relationship-determining modules to identify new relationships and patterns in the data.
  • the system may use these relationships and patterns to create new characteristics, which will be used when evaluating the received data.
  • the system may identify new relationships and patterns in the data.
  • the system may then skip the irrelevant characteristics, increasing performance.
  • the system may also use the relationship-determining module to identify characteristics that are related to each other.
  • the system may group these related characteristics together, as a group of characteristics. Any title may be given to this group of characteristics, or to the group of individuals, data, data sources, or metrics that have a strong relationship with that group of characteristics.
  • the system may use the relationship-determining module to determine the relationships between the groups of characteristics and the characteristics, data, individuals, and the other groups of characteristics. In this manner, personality types may be identified and analyzed.
  • the system may receive a metric, representing an overall goal or interest of a particular organization.
  • the term metric broadly refers to any attribute, measurement, goal, strategy, or other information of interest to an organization.
  • the metric may also consist of a number of sub-metrics.
  • the term sub-metric broadly refers to any attribute, measurement, goal, strategy, or other information related to the metric.
  • the system may use a suitable relationship-determining module to identify the relationship between the metric and the characteristics. In this way, the system may further determine the relationship between the metrics and the individuals.
  • the system may also determine the relationship between groups of characteristics and the metric, and individuals and the metric. In this manner, the organization may gain information on how personality types or individuals contribute to the metric it is interested in.
  • a visualization module i.e., a software component, a hardware component, or a combination of a software component and a hardware component
  • the user may select two areas of interest.
  • the selections may comprise one or more metrics, sub-metrics, characteristics, groups of characteristics, individuals, data items, data sources, or any grouping of the same.
  • the system may use the relationships for those selections to calculate an overall relationship between the two.
  • the system may then represent this overall relationship as a single value or descriptor.
  • the user may assign weights to one or more of the selection items, or change the assigned weights. When the weights are changed, the system may re-calculate all
  • the system may use these weights accordingly when calculating the overall relationship between the selections.
  • the system may also determine the relationships between one selection and the underlying items comprising the other selection.
  • the system may then compute a single value or descriptor for the underlying items. In this manner, the user is able to determine how the underlying items contribute to the overall relationship between the selections.
  • the system may also receive a plurality of threshold criteria.
  • threshold criteria broadly refers to any value, term, event, or descriptor related to one or more data items, data sources, individuals,
  • the threshold criteria may represent a specific event, (e.g., an individual has changed their job description), a keyword (e.g., an advertising keyword was mentioned in a blog post), a value (e.g., a relationship is at, above, or below the criteria), a transaction (e.g., an individual has booked a flight), or any other criteria about which the organization wishes to be informed.
  • the system may output notifications when any threshold criteria are met.
  • FIG. 1 is a block diagram of an exemplary embodiment of a characteristic-based server 100.
  • characteristic-based system 100 may include a network interface 102, a memory module 106, a processing module 104, a visualization module 108, and one or more interconnected information storage units, such as, for example, a characteristic storage unit 110, a relationship storage unit 111, a metric storage unit 112, an individual descriptor storage unit 114, a data item storage unit 116, a threshold criteria storage unit 118, a note storage unit 120, and a group storage unit 122.
  • a characteristic storage unit 110 such as, for example, a characteristic storage unit 110, a relationship storage unit 111, a metric storage unit 112, an individual descriptor storage unit 114, a data item storage unit 116, a threshold criteria storage unit 118, a note storage unit 120, and a group storage unit 122.
  • characteristic- based server 100 may include only one database that would include the data of storage units 110-122.
  • one or more storage units may be separate units, connected to server 100 through network interface 102.
  • Network interface 102 may be one or more devices used to facilitate the transfer of information between server 100 and external components, such as user terminals 140, 142 and data sources 144, 146.
  • Network interface module 102 may receive user requests from local user terminal 140 or remote user terminal 142, and route those requests to processing module 104 or visualization module 108.
  • network interface module 102 may be a wired or wireless interface to a local-area network connecting one or more local user terminals 142 and local data sources 144, or wide-area network such as the internet, connecting one or more remote user terminals 142, or remote data sources 146.
  • Network interface module 102 may allow a plurality of local user terminals 140 and remote user terminals 142 to connect to the system, in order to make selections and receive information, alerts, and visualizations.
  • Network interface module 102 may also allow the system to connect to one or more local data sources 144, on a local-area- network, or remote data sources 146, on one or more remote networks.
  • Memory module 106 may represent one or more non-transitory computer-readable storage devices that maintain information that is used by processing module 104 and/or other components internal and external to
  • memory module 106 may include one or more programs that, when executed by processing module 104, perform one or more processes consistent with embodiments of the present invention. Examples of such processes are described below with respect to FIGS. 1-21. Memory module 106 may also include configuration data that may be used by processing module 104 to present user interface screens and visualizations to user terminals 140 and 142. Examples of such screens are described in greater detail with respect to FIGS. 9-16.
  • Processing module 104 may further include a data collection module 130, a grouping module 124, a pattern recognition module 126, and a relationship analysis module 128.
  • Data collection module 130 may include components for collecting data items from data sources, using network interface 102. As described in more detail below, data items collected by the data collection module may include any information pertaining to an individual.
  • Relationship analysis module 128 may include components for determining the existence and strength of a relationship between two items. For example, and as described in greater detail below, relationship analysis module 128 may include a natural-language processing component for determining the relationship between two items. Relationship analysis module 128 may also include a sentiment analysis component for determining a sentiment associated with the relationship. Grouping module 124 may include components for identifying groups of related items. For example, and as described in greater detail below, grouping module 124 may use relationships identified by relationship analysis module 128 to identify groups of related items. Pattern recognition module 126 may include components for identifying patterns in the received data. For example, and as described in greater detail below, pattern recognition module 126 may include pattern recognition algorithms known in the art to identify new characteristics based on patterns of received data.
  • characteristic-based server 100 may also include a plurality of interconnected storage units, 110-122.
  • server 100 may include a storage unit module (not shown) having components for controlling access to storage units 110-122.
  • a storage unit module may include a query function that, in response to a match request, may query information stored in one or more of storage units 110-122 to identify characteristics, data items, or metrics meeting specified criteria.
  • Storage units 110-122 may be configured using any appropriate type of unit that facilitates the storage of data, as well as the locating, accessing, and retrieving of data stored in the storage units.
  • Characteristic storage unit 110 may store general characteristics of individuals.
  • the term characteristic broadly refers to any attribute, trait, value or other factor associated, whether objectively or subjectively, with an individual or group of individuals.
  • a characteristic may reflect a number of attributes that may be applicable to one or more individuals, such as types of previously or currently held fields of work (e.g., salesperson), professional or personal values (e.g., environmentalism), location (e.g., New York), social interactions (e.g., trendsetter), emotional traits (e.g., generally negative), user- defined characteristics, or others.
  • Relationship storage unit 111 may store information regarding relationships between one or more characteristics, individuals, groups of
  • Relationship storage unit 111 may also store values, weights, or other information calculated by processing module 104 or visualization module 108.
  • Data item storage unit 116 may store data collected by data collection module 130.
  • Data item storage unit 116 may also store metadata associated with the data items, describing the data item. For example, metadata may include the data source the data item was collected from, the time the data item was posted or created, the time the data item was collected, the type of data item (e.g., a blog post), or the individual with which the data is associated.
  • Data item storage unit 116 may also store data items received, or created by characteristic-based server 100.
  • Metric storage unit 112 may store metrics and sub-metrics for an organization.
  • a metric broadly refers to any measurement, criteria, goal, or information of interest to an organization.
  • a given organization may be interested in "brand awareness," or how likely a given person is to recognize the organization's brand.
  • the metric may also be comprised of sub-metrics.
  • a sub-metric refers to any information related to a metric.
  • sub- metrics related to brand awareness may include "internet mentions" for that brand, how widely those mentions are distributed, how the mentions describe the brand, number of sales, or others.
  • Individual descriptor storage unit 114 may store descriptors of specific individuals.
  • an individual descriptor includes any information that identifies a specific individual, as opposed to a group of people. Descriptors may include names, addresses, employee numbers, drivers license numbers, credit card and other banking account information, social security numbers, behavioral profiles, relationship or social network information, linguistic styles or writing, voice recognition, image recognition, or any other unique identifiers. In this manner, each descriptor or group of descriptors may be used to identify a unique individual.
  • Note storage unit 120 may store notes, consisting of information entered by one or more users, which are associated with one or more individual descriptors, groups, relationships, metrics, sub-metrics, data items, or data sources.
  • the information may include textual, graphical, audio, or video information.
  • a user may enter a description of a specific group, as the "treehugger" group. This description may allow users to more easily refer to, and understand the characteristics that comprise that group.
  • Group storage unit 122 may store groups, consisting of a plurality of characteristics, or other groups. These groups may allow users to more easily identify and understand categories of individuals.
  • Visualization module 108 may further include a selection module 132 and a calculation module 134.
  • Selection module 132 may include components for receiving user selections from network interface module 102.
  • selection module 132 may allow users on remote terminals to make selections.
  • User selections may consist of one or more individual descriptors, metrics, sub-metrics, characteristics, groups, data items, data sources, or groups thereof.
  • Calculation module 134 may include components for determining the relationships between the selected groups and the remaining groups, data items, metrics, sub-metrics, characteristics, data sources, and individuals. This may include using the relationships to calculate an overall relationship for a group with respect to the other groups, data items, metrics, characteristics, data sources, and individuals.
  • Calculation module 134 may also receive weights associated with a group, data item, metric, sub-metric, characteristic, data source, or individual, and use the weights in conjunction with the stored relationships when determining the overall relationship for a selection.
  • Visualization module 108 may use the calculated values for a selection to build a screen containing at least one selection, and a
  • Visualization module 108 may also include additional information about the selection in the screen. For example, and as discussed in more detail below, selection module 132 may receive a selection of an individual and a selection of a metric. Calculation module 134 may determine the overall relationship between the individual and metric based on the stored relationships. Visualization module 108 may return a screen containing information about the individual and a single descriptor of the overall relationship.
  • Characteristic server 100 may consist of a single computer or mainframe, containing at least a processor, memory, storage, and a network interface. Server 100 may optionally be implemented as a combination of
  • server 100 may be implemented in a number of different computers, connected to each other either through a local-area network (LAN) or wide-area network (WAN).
  • Data collection module 130 may optionally comprise search engine tools known in the art, operable to find data sources and data items relevant to the search criteria, such as an individual.
  • Storage units 110-122 may comprise any computer-readable medium known in the art, including databases, file systems, or remote servers.
  • FIG. 2 is a flowchart demonstrating an exemplary process 200 for characteristic-based profiling consistent with the present invention.
  • characteristic-based system 100 may use process 200 to determine the relationship between an individual, or groups of individual descriptors, and a user-specified metric based on a number of characteristics.
  • process 200 may begin by receiving a number of characteristics, an individual descriptor, and a metric.
  • a metric broadly refers to any measurement, goal, interest, parameter, or other information that an organization may be interested in learning.
  • the metric will be an overall goal or measurement related to a business.
  • the system uses the data and
  • the system may also be used to identify other factors related to the metric, such as characteristics, groups of characteristics, data sources, or sub-metrics. By recognizing new characteristics as data is processed, the system may also identify new, previously unknown, customers or groups of customers related to the metric. For example, the system may use a pattern recognition module 126, described above, to determine patterns of data that are not defined as characteristics, but which occur on a regular basis. Once recognized, the system may automatically define these patterns as new
  • characteristics broadly refer to any attribute, trait, value, or other factor associated, whether objectively or subjectively, with an individual or group of individuals.
  • characteristic types may comprise: social network (influencer, follower, etc.); sentimental (positive, neutral, etc.);
  • temperamental emotionless, dramatic, etc.
  • attitudinal health conscious, eco- friendly, etc.
  • psychographic personality factors, personality-derived factors, etc.
  • demographic age, gender, etc.
  • transactional past purchases, rewards, etc.
  • a characteristic may be an objective factor, such as age or income, a subjective factor, such as "eco-friendly,” or a combination of objective and subjective factors. These characteristics are typically selected by a user, based on known templates, or on the types of individuals they believe will be relevant to one or more metrics or sub-metrics. Alternatively, or in addition, and as described in more detail below, the system itself may identify characteristics that are relevant to the metric as it analyzes the data items. These characteristics may also be obtained or purchased from other data sources, such as marketing databases, or public websites, discussion boards, or databases.
  • Individual descriptors broadly refer to any information that may be used to identify a specific individual, including account information, license numbers, phone numbers, email addresses, name, relationship information, behavioral profile, nicknames or aliases, or any information that may be used to differentiate one individual from a group. These descriptors may be received from organizations, users, or internal or external data sources, as described below. Further, an individual descriptor may contain multiple pieces of information that collectively identify a specific person. For example, an individual descriptor may consist of a name, driver's license number, credit card account number, and street address, which may be used collectively to identify a specific person. This example is not limiting, and any information that uniquely identifies an individual may be part of an individual descriptor. For another example, an individual descriptor may consist only of social network information, which describes a person by their social or business
  • the system may receive a plurality of data items, characteristics, sub-metrics, an individual descriptor, and a metric.
  • the data items may be received from a plurality of data sources.
  • the system may create relevant data items for the individual. In one embodiment, the system accesses all data sources that may have relevant information about the metric. These data sources may comprise; internal data sources (e.g.
  • the purchased data sources may also contain characteristics, metrics, or individual descriptors.
  • the system may only access data from sources that have been marked as relevant for one or more individual descriptors, metrics, groups, or sub-metrics.
  • data sources may contain both structured, and
  • unstructured data which may be qualitative and subjective, quantitative and objective, or a combination of both.
  • Structured data broadly refers to any data that is placed into a pre-existing structure such as a database, spreadsheet, or form.
  • Unstructured data broadly refers to data that does not have a defined structure, such as prose, news articles, blog posts, comments, messages, emoticons, images, video, audio, or other freely-entered data.
  • Quantitative and objective data broadly concerns factual, measurable subjects. For example, quantitative data may be described in terms of quantity, such as a numerical value or range.
  • qualitative and subjective data broadly describes items in terms of a quality or categorization wherein the quality or category may not be fully defined. For example, qualitative and subjective data may describe objects in terms of warmth and flavor.
  • the system may use an appropriate relationship-determination module (i.e., a software component, a hardware component, or a combination of a software component and a hardware component), utilizing techniques known in the art, to determine the strength of the relationship between the data items and the
  • the system may determine which individuals are associated with the data item.
  • the system may also use other methods to identify the individual associated with, or likely to be associated with a data item.
  • the data item may be associated with a known individual descriptor, such as a username, account, or name.
  • the system may determine when the data item refers to a pseudonym, or includes missing information about an individual. For example, when a data item strongly relates to a known descriptor, but the names do not match, the system may use additional methods to determine whether the two individuals are the same. In such a case, the system may create a pseudonym item, containing a descriptor of the individual associated with the data item. If additional data items are also found to have a strong relationship to both the individual descriptor and the pseudonym, the system may add the information from the pseudonym to the individual descriptor. In this manner, future data items relating to the pseudonym may be identified with the individual. If no strong relationship is found, the system may use the pseudonym to create a new individual descriptor.
  • the system may automatically use the pseudonym to create a new individual descriptor, or add the pseudonym information to an existing individual descriptor, if threshold relationship strengths are met. For example, if the relationship strength between the pseudonym and the descriptor reaches a set value, the system may automatically merge the two. Likewise, if the relationship strength falls below a certain threshold, the system may automatically create a new descriptor based on the pseudonym. This behavior is not limited to names, and the system may perform this action when any of the information in the individual descriptor does not match the information in the data item. In this manner, the system is capable of collecting new information about the individuals, as well as recognizing new individuals.
  • the system creates an association between the data item and the individual descriptor.
  • the system will also mark the data source as relevant to the individual descriptor, so that it may be identified more quickly in the future.
  • the system will next use an appropriate method known in the art, such as, for example, natural language processing, to identify the portions of the data item that are relevant to the individual.
  • the system uses the relevant data portions to create a new data item, containing only the data relevant to one or more individual descriptors.
  • Information atoms broadly refer to any component of information represented in the system, such as data items, metrics, sub-metrics, individual descriptors, characteristics, relationships, sentiments, notes, metadata, or groups thereof.
  • the system may determine a relationship between two information atoms, regardless of whether those information atoms consist of data items, characteristics, or even other relationships. In this manner, the system may flexibly handle new types of information, and analyze relationships that may otherwise go unnoticed.
  • the system uses a suitable relationship-determining module (i.e., a software component, a hardware component, or a combination of a software component and a hardware component) to determine the relationship between the individual descriptors and the characteristics.
  • the relationship- determining module may comprise algorithms known in the art, including one or more of; natural language processing, textual analysis, contextual analysis, direct 1- to-1 mapping, artificial intelligence, image analysis, speech analysis, sentiment analysis, or other suitable techniques known for determining correlations, patterns, sentiments, emotions, temperaments, or relationships.
  • the relationship consists of a magnitude, indicating the strength (or lack thereof) of the relationship, and a direction, indicating whether the relationship is positive, negative, or neutral.
  • the direction simply indicates whether a given relationship represents a positive correlation (i.e. positive direction), a negative correlation (i.e., negative direction), or no correlation (i.e. neutral direction).
  • a positive correlation i.e. positive direction
  • a negative correlation i.e., negative direction
  • no correlation i.e. neutral direction
  • the system may use a relationship- determining technique known in the art to determine the relationship between the data items and the characteristics. This relationship may consist of a magnitude and a direction. The system may also calculate a value for a characteristic based on the relationship between the characteristic and the data item, and the relationship between the data item and the individual descriptor. This is represented in FIG. 3, items 326-332 (first set of relationships) and 334-340. For example, the relationship between individual A 302 and characteristic W 316 will be determined based on second relationship D1W 334 and first relationship D1A 326; where second relationship D1 W 334 represents the relationship between characteristic W and data item 1 310, and first relationship D1A represents the relationship between data item 1 and individual A.
  • the combined relationships will be stored with the characteristics, and associated with the individual descriptor as shown in FIG. 4.
  • the combined scores based on D1A, D1W, 402 to D1A, D1Z 408 are associated with the relationship between individual A 302, and characteristics W 316 to Z 322.
  • the system may also determine the relationship between the characteristics and the metric. This relationship may also consist of a magnitude and direction, as described above.
  • the system may determine this relationship using a suitable relationship-determining module (i.e., a software component, a hardware component, or a combination of a software component and a hardware component), known in the art.
  • FIG. 5 illustrates an example of the third set of relationships determined between characteristics W 316 to Z 322, and metric M 324, represented by MW 342 to MZ 348 respectively.
  • the system may determine the relationship between individual descriptor 302 and metric 324. The system may determine this relationship using a suitable relationship-determining module, as described above.
  • This relationship may also consist of a magnitude and direction, as described above. As shown in FIG. 5, this relationship may be determined based on the relationships between characteristics 316-322 and metric 324, represented by MW 342 to MZ 348, and the relationships between the individual 302 and characteristics 316 to 322, represented as D1A, D1W 402 to D1A, D1Z 408 (the sixth set of relationships).
  • the system may output individual-metric relationship 500, representing the strength of the relationship between individual 302 and metric 324.
  • This score may be represented as a numerical value, a descriptor, an image, or any other means of conveying the overall magnitude and/or direction of the relationship between individual 302 and metric 324.
  • the system may store a "baseline" sentiment for each relationship. For example, when a relationship is first calculated between two items, the system may store the sentiment associated with that relationship as a baseline for the relationship.
  • the sentiment may include any type of meaning, such as a general temperament or mood, an emotion towards a specific thing, or a temporary change in mood.
  • Multiple baselines can be created for an individual based on any number of factors identified as significant, such as context (e.g., with whom the individual is conversing, where the individual is, what did the individual just do or encounter, and what's happening in the world), subject matter, and time.
  • Machine learning, sentiment analysis, natural language processing, image analysis, contextual analysis, voice recognition and speech analysis, and other suitable techniques may be used to identify sentiments, create baselines, and identify the degree of difference from one or more baselines and determine the significance of such difference with respect to the individual, other individuals, and a business.
  • the system may compare the new relationships to one or more of the baseline relationships, in order to determine in what manner and to what extent they differ from one or more of the baselines.
  • the system may determine whether these differences are associated with other data items such as specific topics, interests, events, interactions, situations, noises, or images. These differences may also be associated with a specific time, date, or location.
  • the system may also determine whether the differences are associated with exposure to any other individual, any group of individuals, or any particular media, such as music, television, movies, books, radio, the Internet and specific webpages or a category thereof, mobile applications, advertisements, or promotions.
  • the system may also determine whether the differences are associated with any particular temperature, body temperature, or geographic location.
  • the system may also store these differences. In this manner, the system may calculate and store the changes in each relationship over time.
  • the system may also calculate relationships between each difference and all the other stored items. In this manner, the system may identify how each change in sentiment is related to other relationships, individuals, and other changes in sentiments and business circumstances.
  • a strong relationship between a data item related to one individual that has a strong positive sentiment towards a data item, and differs from that relationship's baseline may be related to other individuals' also expressing strong positive sentiment towards the same item. Or, it may be related to other individuals' expressing mild negative sentiments towards a different data item.
  • the system may also identify patterns in these relationships over time. Likewise, the system may identify patterns in one or more of the baselines over time. The system may also use machine learning or other suitable techniques, as mentioned above, to recognize new baselines in one or more relationships, as more information is collected.
  • system may identify groups of
  • the system may also use a suitable relationship- determining module, as described above, to determine the relationships between the characteristics.
  • the system may identify groups of
  • characteristics W 316, X 318, and Z 322 are strongly related, and the system may group them into group 1 600. Because characteristic Y 320 is not strongly related to the others, the system may not include it in group 1 600.
  • the system may also determine the relationship between the groups and the metric, based on the underlying characteristics. For example, the system may use a suitable relationship-determining module, as described above, to determine the relationships between the groups and the characteristics. For example, as shown in FIG. 7, the system determines a fourth set of relationships G1W 702 to G1Z 708 based on the relationship between group 1 600 and characteristics W 316 to Z 322. As described above, the relationship may contain a magnitude and direction. As shown in FIG.
  • the system may determine the group-metric relationship 800 (one of the fifth set of relationships) between group 1 600 and metric 324 based on the third set of relationship values MW 342 to MZ 348 and the fourth set of relationships G1W 702 to G1Z 708.
  • the system may output group-metric relationship 800, which may be represented as a numerical value, a descriptor, an image, or any other means of conveying the magnitude and/or direction of the relationship.
  • the system may also determine the relationship between the sub-metrics and the metric.
  • the system may also use a suitable relationship-determining module, as described above, to determine the relationships between the groups of characteristics and the sub-metrics.
  • the system may determine a tenth set of relationships, the metric-sub-metric values QM 910 to TM 916 based on the relationship between metric M 324 and sub-metrics Q 902 through T 908.
  • the system may also determine a ninth set of relationships, the group-sub-metric values G1Q 918 through G1T 924, based on the relationships between group 1 600 and sub-metrics Q 902 through T 908.
  • the relationship may contain a magnitude and direction.
  • the system may also determine the overall relationship score for the sub-metrics, based on the group-sub-metric values and metric-sub- metric values. For example, the system may determine an overall relationship for sub-metric Q 902 to metric M 324 based on G1Q 918 and QM 910. The system may output this information, as described above. In this manner, the system may determine which of the sub-metrics have the strongest relationship to the overall metric M 324.
  • the system may also determine an overall score for a metric, representing how successful the company is in meeting its metric, based on the collected data.
  • FIG. 10 shows an example of overall metric score 1000, based on a plurality of metric sub-scores, 1002-1008.
  • the metric sub-scores are determined based on the metric-sub-metric values 910-918, as well as the group scores 800, 1010, 1012 for one or more groups having strong relationships to the sub-metrics.
  • the system may determine score 1000 for the metric based on one or more of the sub-scores 1002-1008. As described above, the system may output this score using a suitable descriptor or value, at step 220.
  • FIG. 11 shows another example of the relationships between groups, characteristics, sub-metrics, and the metric.
  • the system may use a suitable relationship-determining module, as described above, to determine an eighth set of relationship between groups, represented as G12 1102, G13 1104, and G23 1106.
  • the system may identify groups of characteristics that have strong relationships to each other using grouping module 124. In this manner, the system may also create larger groups, in the event that less granularity is desired.
  • a map of relevant data may be built from internal data, in order to identify relevant characteristics and data sources. For example, an organization may already possess information about its customers or relevant individuals. The system may analyze this data, using the steps described above. The system may use pattern recognition module 126 to identify relevant characteristics. Once the internal data has been processed, the system may use these characteristics when analyzing data from external data sources. This may save time and increase performance, since the system will use fewer irrelevant characteristics when analyzing the new data. Additionally, in this manner, the system may use information describing individuals it is interested in, without revealing any of the individuals' descriptors. This is because only characteristics, groups, or other mapped data is used when accessing external data sources.
  • FIG. 12 shows an exemplary embodiment of a visualization screen for an individual.
  • Screen 1200 may comprise an individual descriptor window 1202, a notification window 1204, a note window 1206, a score window 1208, and one or more data source identifiers 1210-1214 and weight selection windows 1216-1220.
  • Individual descriptor window 1202 may contain information describing an individual, based on the individual descriptor for that individual.
  • Notification window 1204 may display any notifications related to the individual.
  • Note window 1206 may display notes related to an individual.
  • Note window 1206 may also allow remote users to enter notes, which will be stored and associated with the individual's descriptor. Thus, the notes related to an individual may be entered by a user, and associated with that user, or available to all users.
  • Score window 1208 may contain the overall score for the user, relative to a metric, as computed above.
  • Source identifiers 1210- 1214 may contain icons, text, or other indicators of data sources that have strong relationships to the individual, as determined above.
  • Weight selection windows 1216- 1220 allow remote users to view the current weights assigned to the data sources. Weight selection windows 1216-1220 may also allow remote users to enter new weights for the data sources, causing visualization module 108 to re-calculate relationships and scores as described above.
  • screen 1200 allows users at remote terminals to view information related to individuals, such as the individual's descriptor, notes, notifications, and score. One or more of these components may be missing, or present in a different quantity, or different positions than shown.
  • FIG. 13 shows an alternative embodiment of a screen related to an individual.
  • Screen 1300 may comprise an individual descriptor window 1302, notifications window 1304, note window 1306, communication options window 1308, and data sources window 1310.
  • Communication options window may contain one more representations of the preferred communications methods for the individual. Preferred communications methods may be determined by frequency of use, stated preferences, or weights assigned by a user. The preferred communications window may also allow a remote user to select a particular one of the preferred communications methods.
  • the system may present the user with a communication screen, allowing the user to enter a message, or otherwise communicate with the individual.
  • a communication screen allowing the user to enter a message, or otherwise communicate with the individual.
  • One or more of these components may be missing, duplicated, or in different positions than shown.
  • FIG. 14 shows exemplary communication screen 1400, allowing a remote user to send a message to the individual.
  • Screen 1400 may comprise an individual descriptor window 1402, notifications window 1404, note window 1406, and message window 1408.
  • Screen 1400 may permit the remote user to enter a message into the message window, or otherwise communicate with the individual.
  • the system may send the message to the individual, using the selected
  • communication medium such as email, text message, voice message, video, or other communication methods.
  • the system may use existing communication methods such as voice chat, video chat, instant messaging, or phone to permit the user to communicate interactively with the individual.
  • voice chat voice chat
  • video chat instant messaging
  • phone phone
  • FIG. 15 shows exemplary notification screen 1500, allowing a remote user to view notifications related to multiple individuals.
  • the screen may comprise multiple individual descriptor windows 1502-1506, and one or more threshold criteria windows 1508-1512.
  • Threshold criteria windows 1508-1512 may describe the criteria or event that caused the notifications to be sent.
  • the threshold criteria windows 1508-1512 may also display one or more data items related to the notification. One or more of these components may be missing, duplicated, or in different positions than shown.
  • FIG. 16 shows exemplary notification screen 1600 for a single individual.
  • the screen may comprise individual descriptor window 1602,
  • communication options window 1604, note window 1606, data source window 1608, threshold criteria window 1610, data item window 1612, and score window 1614 may be missing, duplicated, or in different positions than shown.
  • FIG. 17 is a flowchart demonstrating an exemplary process 1700 for determining relationship baselines consistent with the present invention.
  • characteristic-based system 100 may use process 1700 to determine the baseline for relationships between a number of information atoms.
  • process 1700 may begin at step 1702 by receiving a first set of relationships between information atoms. This is represented in FIG. 18, showing relationships 1810-1820 between information atoms 1800-1806.
  • the system may receive a second set of relationships between the information atoms. These are represented in FIG. 19 as relationships 1910-1920.
  • the system may identify how differences D1 2002 - D6 2022 are related to other information atoms.
  • information atom broadly refers to any component of information represented in the system, including data items, characteristics, individuals, metrics, sub-metrics, relationships, baselines, differences, or groups thereof.
  • FIG. 21 depicts an example identifying how difference D1 2002 is related to information atoms 1800-1806, first relationships 1810-1820, second relationships 1910-1920, baselines B1 2000 - B6 2010, and differences D2 2006 - D6 2014.
  • the system may identify how D1 2002 is related, or not related, to the other information atoms.
  • the system may also identify how the other differences D2 2006 - D6 2014 are related to the other information atoms.
  • systems and methods consistent with the invention provide a characteristic-based system that allows an organization to identify, organize, describe, and visualize the relationships between individual descriptors, characteristics, and metrics.
  • certain aspects and embodiments are described herein with reference to the components illustrated in FIGS. 1-21.
  • the functionality of the illustrated components may overlap, however, and may be present in a fewer or greater number of elements and components. Further, all or part of the functionality of the illustrated elements may co-exist or be distributed among several geographically dispersed locations.
  • each "module” may be embodied as a software component, a hardware component, or a combination of a software component and a hardware component.
  • embodiments, features, aspects and principles of the present invention may be implemented in various environments and are not limited to the illustrated environments.
  • FIGS. 1-21 are exemplary and not intended to be limiting. Thus, other process stages may be used, and even with the processes depicted in FIGS. 1-21 , the particular order of events may vary without departing from the scope of the present invention. Moreover, certain process stages may not be present and additional stages may be
  • FIGS 1-21 are implemented in FIGS 1-21. Also, the processes described herein are not inherently related to any particular system or apparatus and may be implemented by any suitable combination of components.

Abstract

La présente invention concerne des systèmes, des appareils et des procédés de détermination du seuil sentimental d'un individu sur la base de plusieurs éléments de données et caractéristiques. Selon l'invention, les éléments de données peuvent comprendre des données objectives et quantitatives, de même que des données subjectives et qualitatives. Le système, l'appareil ou le procédé peuvent obtenir un certain nombre de relations entre des atomes d'informations, identifier des sentiments associés aux relations, et calculer des seuils sentimentaux de ces relations. Les différences par rapport à un quelconque seuil peuvent également être calculées pour déterminer des changements réels de sentiment. Les relations entre ces différences et d'autres données ou relations peuvent également être calculées pour déterminer la façon dont un changement de sentiment est associé à d'autres changements de comportement. Par exemple, on peut déterminer des relations entre une différence particulière et des changements de quelconques paramètre, sous-paramètre, groupe de caractéristiques, élément de données, source de données, caractéristique, sentiment, ou groupes de ceux-ci. Ces relations peuvent également être utilisées pour prédire un comportement ou un sentiment futurs.
PCT/US2013/025070 2012-02-08 2013-02-07 Outils et procédés de détermination d'index de relations sémantiques WO2013119762A1 (fr)

Applications Claiming Priority (14)

Application Number Priority Date Filing Date Title
US201261633246P 2012-02-08 2012-02-08
US61/633,246 2012-02-08
US13/461,670 2012-05-01
US13/461,670 US8341101B1 (en) 2012-02-08 2012-05-01 Determining relationships between data items and individuals, and dynamically calculating a metric score based on groups of characteristics
US13/492,528 US8943004B2 (en) 2012-02-08 2012-06-08 Tools and methods for determining relationship values
US13/492,528 2012-06-08
US13/621,717 US8478702B1 (en) 2012-02-08 2012-09-17 Tools and methods for determining semantic relationship indexes
US13/621,717 2012-09-17
US13/649,452 2012-10-11
US13/649,452 US20130204822A1 (en) 2012-02-08 2012-10-11 Tools and methods for determining relationship values
US201261715415P 2012-10-18 2012-10-18
US61/715,415 2012-10-18
US13/674,692 US20130204823A1 (en) 2012-02-08 2012-11-12 Tools and methods for determining relationship values
US13/674,692 2012-11-12

Publications (1)

Publication Number Publication Date
WO2013119762A1 true WO2013119762A1 (fr) 2013-08-15

Family

ID=48947906

Family Applications (4)

Application Number Title Priority Date Filing Date
PCT/US2013/024119 WO2013119452A1 (fr) 2012-02-08 2013-01-31 Outils et procédés de détermination de valeurs de relation
PCT/US2013/024060 WO2013119451A2 (fr) 2012-02-08 2013-01-31 Outils et procédés pour la détermination de valeurs de relation
PCT/US2013/025128 WO2013119798A1 (fr) 2012-02-08 2013-02-07 Outils et procédés de détermination de valeurs de relation
PCT/US2013/025070 WO2013119762A1 (fr) 2012-02-08 2013-02-07 Outils et procédés de détermination d'index de relations sémantiques

Family Applications Before (3)

Application Number Title Priority Date Filing Date
PCT/US2013/024119 WO2013119452A1 (fr) 2012-02-08 2013-01-31 Outils et procédés de détermination de valeurs de relation
PCT/US2013/024060 WO2013119451A2 (fr) 2012-02-08 2013-01-31 Outils et procédés pour la détermination de valeurs de relation
PCT/US2013/025128 WO2013119798A1 (fr) 2012-02-08 2013-02-07 Outils et procédés de détermination de valeurs de relation

Country Status (3)

Country Link
EP (1) EP2812857A4 (fr)
HK (1) HK1205321A1 (fr)
WO (4) WO2013119452A1 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106062730A (zh) * 2013-09-19 2016-10-26 西斯摩斯公司 用于主动构成内容以便在连续社交通信中使用的系统和方法
IT202100022802A1 (it) * 2021-09-03 2023-03-03 Imasterminds S R L Sistema per ottimizzare la realizzazione di un gruppo di lavoro e relativo metodo.

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070239517A1 (en) * 2006-03-29 2007-10-11 Chung Christina Y Generating a degree of interest in user profile scores in a behavioral targeting system
US20080162574A1 (en) * 2006-11-22 2008-07-03 Sheldon Gilbert Analytical E-Commerce Processing System And Methods
US20100153453A1 (en) * 2007-06-27 2010-06-17 Karen Knowles Enterprises Pty Ltd Communication method, system and products
US20110153377A1 (en) * 2009-12-23 2011-06-23 Victor Novikov Mixing and Targeting Content Types/Items for Users to Promote Optimization Goals

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050108043A1 (en) * 2003-11-17 2005-05-19 Davidson William A. System and method for creating, managing, evaluating, optimizing, business partnership standards and knowledge
US8010460B2 (en) * 2004-09-02 2011-08-30 Linkedin Corporation Method and system for reputation evaluation of online users in a social networking scheme
US8311888B2 (en) * 2005-09-14 2012-11-13 Jumptap, Inc. Revenue models associated with syndication of a behavioral profile using a monetization platform
US7908237B2 (en) * 2007-06-29 2011-03-15 International Business Machines Corporation Method and apparatus for identifying unexpected behavior of a customer in a retail environment using detected location data, temperature, humidity, lighting conditions, music, and odors
US20090132366A1 (en) * 2007-11-15 2009-05-21 Microsoft Corporation Recognizing and crediting offline realization of online behavior
US20090216563A1 (en) * 2008-02-25 2009-08-27 Michael Sandoval Electronic profile development, storage, use and systems for taking action based thereon
US20100257028A1 (en) * 2009-04-02 2010-10-07 Talk3, Inc. Methods and systems for extracting and managing latent social networks for use in commercial activities
US9342835B2 (en) * 2009-10-09 2016-05-17 Visa U.S.A Systems and methods to deliver targeted advertisements to audience

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070239517A1 (en) * 2006-03-29 2007-10-11 Chung Christina Y Generating a degree of interest in user profile scores in a behavioral targeting system
US20080162574A1 (en) * 2006-11-22 2008-07-03 Sheldon Gilbert Analytical E-Commerce Processing System And Methods
US20100153453A1 (en) * 2007-06-27 2010-06-17 Karen Knowles Enterprises Pty Ltd Communication method, system and products
US20110153377A1 (en) * 2009-12-23 2011-06-23 Victor Novikov Mixing and Targeting Content Types/Items for Users to Promote Optimization Goals

Also Published As

Publication number Publication date
WO2013119451A3 (fr) 2015-06-18
EP2812857A1 (fr) 2014-12-17
EP2812857A4 (fr) 2015-11-04
WO2013119798A1 (fr) 2013-08-15
HK1205321A1 (en) 2015-12-11
WO2013119451A2 (fr) 2013-08-15
WO2013119452A1 (fr) 2013-08-15

Similar Documents

Publication Publication Date Title
US8341101B1 (en) Determining relationships between data items and individuals, and dynamically calculating a metric score based on groups of characteristics
US8943004B2 (en) Tools and methods for determining relationship values
Ibrahim et al. A text analytics approach for online retailing service improvement: Evidence from Twitter
Eslami et al. Which online reviews do consumers find most helpful? A multi-method investigation
US11100523B2 (en) Determining relationship values
Gensler et al. Listen to your customers: Insights into brand image using online consumer-generated product reviews
Yan et al. Customer revisit intention to restaurants: Evidence from online reviews
US10565602B1 (en) Method and system for obtaining leads based on data derived from a variety of sources
US9799035B2 (en) Customer feedback analyzer
Sohail et al. Feature extraction and analysis of online reviews for the recommendation of books using opinion mining technique
Lu et al. BizSeeker: a hybrid semantic recommendation system for personalized government‐to‐business e‐services
US20060161553A1 (en) Systems and methods for providing user interaction based profiles
Li et al. Dynamic mapping of design elements and affective responses: a machine learning based method for affective design
US20130204823A1 (en) Tools and methods for determining relationship values
US8478702B1 (en) Tools and methods for determining semantic relationship indexes
JP2013239160A (ja) 情報提供システム、情報提供方法、及び情報提供プログラム
Semerádová et al. Computer estimation of customer similarity with Facebook lookalikes: Advantages and disadvantages of hyper-targeting
Xia et al. Can online rating reflect authentic customer purchase feelings? Understanding how customer dissatisfaction relates to negative reviews
Hollebeek et al. Engaging consumers through artificially intelligent technologies: Systematic review, conceptual model, and further research
Lappeman et al. Exploring online sentiment (OS) as a measure of customer experience (CX) for telecommunication services
Sysko-Romańczuk et al. Data modalities, consumer attributes and recommendation performance in the fashion industry
Aghdaie et al. Customer-oriented benefit segmentation: an integrated approach
WO2013119762A1 (fr) Outils et procédés de détermination d'index de relations sémantiques
Sun et al. Online communities and offline sales: considerations on visiting behavior dimensions and online community types
Ping et al. Enhanced review facilitation service for C2C support: Machine learning approaches

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13746440

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13746440

Country of ref document: EP

Kind code of ref document: A1