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
  • G06Q30/0201—Market modelling; Market analysis; Collecting market data

Definitions

• the present invention relates generally to computer-based " decision systems and methods and, more particularly,, to a system and method for solving problems n multiple domains or subject matter areas on the web/internet,
• Web-based decision aids are intended for use in human problem-solving, speciiieally to assist end-users - including consumers, professional intermediaries, technicians and other professional personnel - in decision-making within subject matters selected by the user.
• An example of a decision aid is presented in U.S. Patent No. 7,761,345 to Marin et a.l.
• subsets of criteria represent Selection Criteria that form the foundation for a. user's Overall Stated Importance for each category. While Overall Stated importance allows users to designate personal importance of each category by setting specific criteria to Not Important, Less Important or More important, it is the relationship between Selection Criteria and Overall Stated Importance thai is paramount.
• the inteliigeni agent provides insight into the user's thought process through semantic processing, thereby teasing out inconsistencies as well as similarities within categories, across categories and across topics. Furthermore, by knowing what other users have done within a similar- domain, the intelligent agent can provide instant feedback to a user newly exploring a domain, for example:
• users may not know the result they Want or Need, or users may kno they Want or Need a category of result but may be unclear as to the nuances as well as breadth of the various options; additionally, users may be initially predisposed toward a type of result that, upon introspection and coaching, is exposed as adverse to their Wants and Needs.
• decisiomng within the system goes beyond effectuating a purchase or result that a user already knows that they want.
• the system recognizes that both Wants and Needs drive behavior, and that oftentimes Needs eclipse Wants, as for instance when Needs pertain to medical requirements, or to financial or time constraints.
• the system .contains logic such that Needs, trump Wants; specifically, an algorithm devalues potential results by assigning a higher weight to Needs than to Wants (note that an exception to this rule occurs when the Overall Stated Importance for a Need is less than for a Want).
• Coaching by the inteiiigent agent transcends decision-making into outcomes of self satisfaction, self fulfillment, self awareness or success; within such an environment, results go beyond tangible products and services, such that they may include pursuing an entirely different course of action.
• each returned provider varies in degree-of-fit of the same conceptual product (for instance, a Daycare facility, -a -Senio Living Facility, a University). Via percentage-based calculations, Single-Product providers are designated as Most Relevant, Relevant and Of Interest. Within the Multi-Product Method, returned providers, vary in degree- of-fit of potentially many products (for instance, Financial Investment. Alternatives, Preventive Health Measures, End-of-life Considerations).
• a provider is not penalized for providing more products and services than a user ' s specified Wants- and Needs; rather, a provider is penalized only for providing less products and services; as such, in the extreme case, a user who selects zero criteria, whereby no Wants and Needs are specified, will be returned all providers as Most Relevant; however, as the user selects more criteria, more providers are disqualified as Most Relevant and Relevant providers.
• This inverse relationship can be represented as follows:
• an Overall Stated Importance of More important calculated for a Need category acts to devalue an provider whose corresponding attributes have a -less-than favorable match (determined via a. percentage-based calculation) regardless of isolated criteria producing a favorable match, or entire Want categories producing an overall highly favorable match. In such cases. Needs trump Wants, and a provider original iy deemed Most Relevant is devalued to Relevant; a provider originally deemed Relevant is devalued to Less Relevant.
• the results container for the Multi-Product method does not point to a list of singular providers; rather, it evaluates every day Wants and Needs then translates those into potential products and services. Rather than present a pre-defined, static set of results for a single provider, then, apply user preference to narrow the choice, the system purposefully separaies providers' products and services from the user's field of choices, such that user selected criteria are evaluated to return results relevant to the user's Selection Criteria and Overall Stated importance.
• the system harnesses technology to reduce the field of results to those with the highest degree of user-relevancy prior to presenting the set to the user, such that a decision is made from a customized solution set, containing only elements that have been pre-quaiifkd as being applicable to the user's Wants and Needs. Only after the user chooses a product or service to review does a list of qualified providers appear, each containing -a web Internet link/URL leading to provider-specific products or services.
• Fig, 1 illustrates an. embodiment of the system of the invention where users connect through deskto computers, laptops, tablets and mobile devices to web servers, thereby gaining access to Ul, algorithms and databases of the system and, through the system, to third party server and. content;
• FIG. 2 through Fig, 6 are examples of User inter ace renderings in an embodiment -of the invention
• Fig. 513 also shows .an embodiment of logic used to manifest content display:
• Fig. 7 lists ⁇ seif eam g behaviors intrinsic to an embodiment of the present invention., and that enable the system to filter results based on user demonstrated behaviors, as well as Persona-type behaviors, which act as baselines from which to compare users and determine relevancy;
• Fig. 8 displays a graphic and a table; the graphic represents a visual cue, presented for end-user benefit, that summarizes user-stated level, of importance; the table represents test cases for calculating Overall Stated importance values;
• Fig. 9 illustrates mapping within the Muiti -Product method, structured as a semantic network
• Figs. 10 through 12 display instances of ontologieal relationships contained within the Multi-Product Method
• FIGs. 13 through 17 display flow diagrams, use cases -and instances of conditional feedback tied to an embodiment of the present invention
• Fig. 18 is an illustration of a mathematical, integer-based algorithm that can be used to represent user .choices in a manner, by which choices can be electronical iy compared to those of other users, and for the end-purpose of connecting users otherwise unknown to one another;
• Fig. 19 through 22 are examples- of logic, tests & conditions which assess user input and drive intelligent feedback and display of results within an embodiment of the current invention
• Fig. 23 is a flow chart depicting of the use action of accessing anew (adding) or updating a decision aid, which upon save, is displayed and stored within their member view of ihe website;
• Fig. 24 is a flow chart depicting, the user action of accessing a saved decision aid. with the purpose of i) viewing displayed results ' and ii) being presented with options of connecting with other users based on the user's specified Social Share criteria, and with similarity to -other users as evaluated via the Social Connect method;
• FIG. 25 is -a flow chart illustrating providers associated with user Stated Importance and Selection Criteria,, based on execution of either the Single-Product method or the Multi-Product method;
• Fig, 26 is a flow chart illustrating method, for displaying a social blog, expertise or agent feedback, based on user request and the format of the associated content:
• Fig. .27 is a flow chart showing adding, displaying and -editing user profile information, which is used to permit saving of decision aids and connecting with other users;
• Pig. 28 is a flow chart illustrating a method for determining the visual display elements appropriate to disclose to other users, based on the originating user's specified Social Share criteria;
• Fig. 29 is a flow chart illustrating a structured process for creating proprietary Content Input Forms which are tied to decision, aids by subject matter and which capture products and services from either Single-Product or Multi-Product Pro viders;
• FIG. 30 i a flow chart showing a method for displaying, via intelligent agent recommendations, content of higher-order relevancy from a proprietary database, which is based on criteria either within a. user-specified decision aid, across user decision aids, or across users who have demonstrated similarity in preferences;
• Fig. 31 is a -flow chart showing a method for determining -the visual display of icons for the decision aids, based on updates to the decision aid by users, by providers or by the intelligent agent;
• Fig. 32 is a flow chart illustrating a method for determining compatibility or percentage similarity of users' decision aids:
• Figs. 33.a-3.3f present an example walkthrough of user interfaces in an -embodiment of the present invention that are common across Single-Product and Multi-Product Domains, and where Fig. 33-b, 33-c, 33-d and 33-e provide an embodiment of the Single-Product Domain, specifically i) selecting criteria and level of importance, ii) viewing Intelligent Agent feedback, iii) accessing social postings, and iv) reviewing relevant results,
• a web-based decision aid for use across domains to assist in human problem solving.
• Advantages of embodiments of the present invention include, without limitation, assisting the user in making decisions through intelligent agent expertise, as well as through related eComrnerce, social networking, guided content search and delivery of context-rich content,
• web profiles typically define members in their entirety, such that preferences are tied to a profile as a whole
• embodiments of the present invention delineates a member's preferences based on specific subject matter areas, assessing those preferences by implicit and explicit user behaviors.
• any decisio aid can be easily added, deleted or updated, without disrupting or altering social interactions, eCommeree and search capabilities related to a member's other decision aids.
• Fig. 1 illustrates an embodiment of the system of the invention.
• Users via browser- enabled controls available on remote access devices such as smart phones, tablets and personal computers (1101 through 1 103), connect -to one or more web servers (1109) containing an embodiment of the invention's algorithms, data schemes and databases, through load balancing and clustering technology shown in 1104 and user and provider portal cluster or internet 110:5,
• 1106 and 1 107 represent access to the system's Content Management System 1 108, enabling creation of proprietary content, as well as management of links to third-party products, services and content deemed relevant to the user, as determined by ontoJogical subsets and percentage-similarity calculations described, in Figs. 9- 12 and Figs. 19-22, discussed below.
• Figs. 2 A through 4B there are shown wireframes and renderings representing screens in an embodiment of the present invention's user interface.
• the graphical displays are shown to convey primary interfaces - namely Add/Update Profile & Domains, Choose Wants & Needs and View Results - not .as limitations of the user web- experience, which includes overlay windows, mouse-over tool tips and browser windows,
• a wireframe for Add/Update- Pro file & Domains is displayed, where 1201 permits user -access to previously saved decision aids ⁇ those for which Wants and Needs have been chosen by the use -earlier, either during the current session or a prior session.
• 1201 permits user -access to previously saved decision aids ⁇ those for which Wants and Needs have been chosen by the use -earlier, either during the current session or a prior session.
• 1 Within the container 1201 , as with other containers within the system, are reveal/hide controls, a scroll bar to access content beyond that displayed within the boundaries of the container, and active-click labels used to launch content.
• the container represented by .1202 indicates a container through which a user may choose from decision aids to add. to their member account; specifically, the container lists al l possible decision aids, grouped by topical neighborhood , The container represented by 1203 permits user access to Other's Decisio Aids; this container lists all decision aids created by other members who have granted the user permission to review and comment on their decision aids.
• 1204 indicates a Sign-in / Sign-out capability
• 1205 indicates a container for adding or updating a user profile, for which- example mandatory fields include Unique User Name and Password, Verified email Address and User birth Year.
• the My Connections container represented by 1206 permits users to establish a master list of family members and friends, from which the user later grants review & comment permissions for decision aids, determined by the user on a ease-by-case basis versus en masse.
• My Connections are initially established by specifying an email address contained within a member database; in cases where an email address oasts, a connectio confirmation message is displayed, along with associated User Name.
• the list of all those members within a user's My Connections list is displayed on the user-specific view of the homepage.
• 1207 indicates Help, About and Terms & Conditions links
• inset 1208 is a graphical rendering of the wireframe after a user has created an- account and added decision aids. An -enlarged view of graphical rendering is- provided as Fig. 2B.
• Fig. 3 illustrates a graphical rendering of Choose Wants & Needs, where 13-01 enables a user to name the decision aid, and where 1302 represents a container for capturing high-level filter information.
• user-demographics are used to narrow resultant datasets that are retrieved from the web/Internet. Beyond cross-domain filters such as birth year, the value of which is captured at the profile level, certain filters are domain-specific, or arc relevant i multiple but not all domains.
• Location for example, is necessar within a Financial Assets domain to parse-out banks thai do not operate in certain U.S. states, This filter isolates location- relevant providers from the universe of all providers identified as potentially relevant, thereby preventing the display of non-meaningful banks. Location is also applicable to domains returning relevant facilities (for example, Senior Living, Golf Courses); in such cases. Location narrows the universe of providers to a subset satisfying a true condition of user-chosen geo- centricity.
• third-party tools are used to streamline and/or expand the Location example.
• the user interface allows users to enter city/state versus zip code; here, third-party apps are leveraged to convert, city/state to zip code.
• the user interface allows users to enter a search radius from the specified city/state or zip code. Again, third-party apps determine all zip codes within the specified radius; that series of zip codes is then used to retrieve the subset of relevant providers.
• results for the decision aid are filtered through an Exception Rule Base; the same Exception Rule Base i applied to Single-Product Domains and to Multi-Product Domains, For example, birth year is relevant for domains such as Senior Living and Healthcare, among others.
• Exception Rule Base i applied to Single-Product Domains and to Multi-Product Domains, For example, birth year is relevant for domains such as Senior Living and Healthcare, among others.
• an intended Senior Living Facility search is for an aged senior, 85 years or older. If the user selects a Want (hat could return a result more suitable for a younger, more active senior, then results are tempered to reflect the birth year filter.
• a facility tagged as an Active Adult Community, and that without the ilter would otherwise be displayed as Most Relevant, is discounted to Relevant; a facility otherwise displayed as Relevant is discounted to Of interest..
• Exception Rule for discounting results can be understood as follows: for any product calculated as Most Relevant and that contains potential exception content, assess a rule base [Ri -.R ⁇ l for limits to the exception content. For example, if Rule K states ''Active Adult Community" ⁇ ⁇ ; 8 Years" for birth year, and if the exception test fails, then discount the product from Most Relevant to Relevant, or from Relevant to Of Interest. The discounting algorithm is not applied if the product is originally returned as Of Interest, since no tier exists beneath Of Interest. [0051 J Now, still referring to the. screen shown in Fig.
• 1303 indicates a container from which the user grants review & comment permission to family members and friends, specifically for this decision aid, regardless of permissions granted for other decision aids.
• a user follows a two-step process, where establishing and managin My Connections, described in Fig. 2A, item 1206, are. treated separately from inviting others to view specific decision aids. Specifically, having a member within My Connections does not in.
• Stated Importance values of Less Important and .More Important are treated equally -namely as a Hag that the criteria/goal has been selected b the user.
• Stated Importance values are used for Intelligent Agent feedback and as reminders/cues to the user, versus weighing results providers.
• Multi-Product Domains results are weighed based on recurrent instances of products satisfyin user-specified goals.
• the system addresses common challenges of survey validity, including (i) low incentive to complete, (ii) perceived risk, of responding truthfully, (Hi) ego inflation, (iv) misinterpretation of queries, (y) Jack of qualifying feedback during the survey, and (vi) mid-point non-committal response to requested Level of Importance.
• selected criteria confront these challenges directly, thereby dramatically improving reliability of user responses, and therefore of result relevancy.
• the. screen of Fig. 3. at 1306 contains further instructions for saving user choices, then viewing intelligent Agent feedback and recommended results.
• Figs. 4A and 4B shows a graphical rendering of View Results, where 1401 enables a user to toggle between Blog containing Intelligent Agent feedback, and social postings from family members and friends; however, if in Fig. 3, item 1303 the user chose to designate the decision aid as private, the social postings tab in 1401 is void of postings—ts. sole contents are instructions- for granting review and comment permissions, via item 1303.
• the Intelligent Agent's Blog contains no Relevancy headers, such as those used to display product results. -Specifically, no mathematical calculation is used to sequence the display of Intelligent Agent feedback; rather, the display bears a 1 : 1 relationship to the sequencing of user-state ' d goals, which is static within an embodiment of the present invention, and is determined by content authors when goals are authored. In cases wher two goals produce the same Intelligent Agent feedback, only .the- higher occurrenc is displayed (this prevents a specific statement from being redundantly displayed in the same container, at the same time).
• 1402 displays criteria that the user selected, as described in Fig. 3, item 1304. Only chosen criteria are displayed, not their associated level of importance.
• the containers within 1 03 and 1404 (Fig. 4.4 ⁇ indicate results of Multi-Product Domains,
• the containers within 1405 and 1406 (Fig. 4B) indicate results of Single-Product Domains, Specifically. 1404 displays whenever a user clicks on a product label within 1403; and 1.406 displays whenever a user clicks on a provider label within 1405. In this manner, intelligent Agent feedback and detailed result descriptions are extracted based on user action.
• a text container indicating the output of the system ' s Internet/web Search String Generator.
• This generator produces ontological representations for ail products that the system ' s Intelligent Agent recommends.
• the system displays a separate search string (via a drop down box) such that each search .string contains the product name, plus the ontological representation of that goal
• the search string which, is generated automatically by utilizing the system's ontological schema, may be overridden via a content author's manual edits,
• a relevant product is 529 Plan (Product: P)
• 529 Plan has only one user-selected goal which triggered its placement within the relevant products subset
• a single search string is displayed, its contents dependent on the ontological representation of the user-selected goal not of the product-— for which an instance may be:
• a relevant products (RP) subset containing three products (Pj, *, P; thai are all ontologically codified as Class M (C ), will give 3 ⁇ 4 a count of 3: within the same relevant products subset, two products (3 ⁇ 4, Pj) ontologically codified as Class N (CN). will give C a count of 2.
• CM along with product family members Pj, Pj and P will precede CN and its product family members P 4 and P 5 .
• product family occurrences translate into degree of relevancy for each product family. When, applied to the interface rendering within Fig. 5 A, these class and product designations become;
• real estate limitations may restrict the display of products per family, resulting in the need for a user control such as the arrow next to Roih 401(h) in 1502—such user controls indicate to the user that additional products, in this case P 3 , : are also placed in the results basket for the product family called Retirement Funds ⁇ H ms. in further explanation, still considering the relevant product subset containing the following five products:
• the esults engine continues to count products within their product families, then displays produei compete in descending order of occurrence.
• administratively-set thresholds may be used to further designate whether a product family is displayed beneath the Most Relevant or Relevant label. For instance, within Multi- roduct Domains, an embodiment of the present invention sets the number of Most Relevant families at a default value of five (5) families. In instances where no more than 5 families are relevant, then no Relevant label is displayed.
• Relevant Products Is the subset of products; identified by intersecting the Potential Relevant Products set (PRP) with user-state goals [00721 2. TRUE, refers to the number times a particular product is identified via the PRP and user-stated goals intersection
• an intelligent Agent feedback tied directly to products is displayed along with the product and therefore is tied to the relevant products (RP) subset intersection, with user-stated goals, which dictates relevancy display for products.
• RP relevant products
• a single instance of Intelligent Agent feedback is used for multiple products, then it is displayed for each, of those products; within the system, this repeated display is not considered redundant, since it can never appear visually in the user interface at the exact same time, in further detail, still referring to Fig.
• each self-learning method becomes an additional Internet/net search results filter.
• the results routines allow for filters to be added, removed or modified without changes to untouched -filters *
• Adaptive algorithms within he system are based on action and inaction of each user, as well a representative behavior (Persona-type X) for each user.
• the outcomes of these algorithms are refined Intelligent Agent suggestions (that is, a reduced set of statements, products & .services deemed relevant to the user).
• Algorithmic categories applied to the system's Intelligent Agent include those shown in Fig. 7, such as the agent's ability to learn to (i) ignore, (ii) append, (iii) befriend, (by.) suggest, (y) refine and (vi) be current,
• Adaptive & non-adaptive methods are applied to the system's neural network engine.
• Non-adaptive methods drive the neural network engine's co nectionistic behaviors-—that is, the mapping of inputs to relevant results.
• the adaptive methods focus on Persona-types assigned to like-users, and are implemented via a threshold-assigned similarity in profile attributes and stated Wants and Needs. Each. Persona-type then filters the displayed results, thereby refining results to reflect behaviors captured within the Persona-type.
• an embodiment of the present, invention enables business-rules to dictate when the neural network engine performs self- adjustments, whereby for any filter Fj the input and/or output nodes of the neural network, assimilate the properties defined by Fj .
• This progression of i) non-adaptive, to ii) adaptive via filters,, to iii ⁇ fully self-adjusting, enables business-driven thresholds and rules to dictate modification to the neural network, rather than granting it organic behavior.
• Semi-automatic updates allo tmnsformation of the network, without encountering ill consequences inherent in. automatic updates being contemplated for Ai systems.
• filtering is a function of the frequency of CN ( ⁇ of CM clicks enacted by a Persona-type) as compared with frequency of each related concept within the ⁇ ci- Cj3 ⁇ 4;.. Filtering within the system also occurs at the Product Family level, where Product Family is synonymous with the construct Class, allowing removal, demotion or promotion of a Product Family results based on frequency of Persona- type access. Multi-Product Domain results, beyond being based on an oi tological match to user queries, are affected by any Persona- type discounting/promoting methods.
• Implicit cues include i) selection of criteria under Wants & Needs, and ii) clicks on Intelligent Agent feedback. Links and products. Explicit cues take the form of "'interest/satisfaction'' checkboxes labeled -Save and— Share.
• the system's Save checkbox is located next to. each, intelligent Agent feedback and resulting product. Whenever a user checks ⁇ Save, a link to that content appears in the— Save container,, facilitating future retrieval, review & action by the user.
• The— Save container is available via the home page ⁇ Figs. 2 and 33-f), displaying links to all content that a user has flagged for future review regardless of whether the user has modified Wants and Need and possibly removed the content from coaching and results containers for a specific domain,
• the present system's Share checkbox is located next to each product. Whenever a user checks -Share, link to tha content appears in the user's ''Postings" blog, with an automatic statement asking family & friends to review the product and provide comments. Family & friends can click the link within the blog to -review the product, which may no longer appear in the results container if the user has modified Wants & Needs. When family & friends click the link, data analytics regarding usage/review/access are updated.
• FIG. 8 summarizes, per Want category and Need category, user- stated level of importance.
• An algorithm applied to underlying criteri is used to generate a visual cue indicating whether the user weighed the category higher or lower in overall importance; effectively, the three-positio control, for user-stated level of importance results in a visual reminder to users and their invited friends & family, as to what's important to the user.
• a table displays test cases for the algorithm, an embodiment of which appears in Pig. 19.
• Fig. 9 visualizes output of the Multi-Product Domain method, wherein a neural network utilizes a proprietary Is-a" based ontology that, although exact in structure across subject areas, contains unique identifiers for each domain, thereby enabling th mapping of user selected criteria (1901 and 1902) chosen from eight categories of Wants and Needs, to related products and services (1903, 1904 and 1905).
• the system's neural network engine functions offline. Although its output is necessary to generate relevant results for users, the engine itself doe not run during a user session.
• the file produced by the neural network engine, which is then accessed runtime to determine user results, is an MxN matrix of Potential Relevant Products (PRP), where M represents the total number of products within the universe u, and N is . the total number of criteria (or goals) within the universe- ⁇ 3 ⁇ 4, such that each column corresponds to a goal.
• PRP Potential Relevant Products
• M represents the total number of products within the universe u
• N is . the total number of criteria (or goals) within the universe- ⁇ 3 ⁇ 4, such that each column corresponds to a goal.
• the value in cell PjGs is true when the Nxl matri for Pi multiplied by the Nxl matrix for Gs equals the Nx l matrix for G5.
• PRP is an extraction of products that are ontologjcally aligned with goals; this extraction method is used within Multi-Product Domains t produce the PRP, and within Single-Product Domains to identify Potential Market Links (PML).
• the first row within Fig 10 indicates PRP ske f execution timing and purpose. Beyond being the set from which relevant products are calculated for each user, the MxN matrix, of Potential Relevant Products is utilized by content authors to write Intelligent Agent feedback based on Goal-to-Product relationships defined within the MxN matrix. Still referring to Fig.
• the PRP solution set is produced using the same method that calculates the Potential Market Links (PML) solution set; the PML is required by Single-Product Domains to determine which goals map to which Market Links (consumer retail products).
• the third row within Fig, 10 indicates PML size, execution timing and purpose.
• Relevant Products (RP) is a subset of the PRP solution set, and is calculated runtime based on the intersection of goals selected b a user, and those products within the PRP thai are associated with user-selected goals.
• RP i produced using the same method that calculates the Relevant Market Links (RML) subset within Single-Product Domains, used to determine which Market Links within the PML solution set are associated with user-selected goals.
• RML Relevant Market Links
• RR Relevant Results
• the inference algorithm that drives neural network engine mapping is used solely to ma goals to products; however, separate methods are used to execute intra-mapping of products and goals.
• P2P Product-to-Product mapping-
• I xN binary matrix multiplier method fails to produce real- world results: while this method can produce a subset of results mathematically, its outcome inadequately reflects relationships produced through human cognition.
• the system uses a derivative method of the one that generates Single-Product results.
• this method calculates "percentage fit.” This P2P capability becomes an important machine self-learning opportunity: enacting this method allows the Intelligen Agent to comment on related products, and to return "near products” that Goal-io-Produci mapping does not return. Beyond the machine self-learning benefit, this method reduces the need for content authors to create additional (though minimally dissimilar) goals to point to products not currently returned via the MxN matrix. Using this method, the system's Intelligent Agent is essentially saying, "The similarity of this product, is so strong to another product I've suggested, that ⁇ am including it with your results '
• Goal-to-Goai mapping within the current embodiment, of the invention, utilizes a method supplemented by human assistance; for any goal chosen b a content author within the system's Content Management System, all other goals that contain, the same class but different subclasses are displayed to the author. This displa is an indication to content authors that goals may be contradictory, or that intelligent Agent feedback, should be written to clarify (dis)similarity of the goals.
• the intelligent Agent Rather than allow the inference engine to remove products of a contradictory nature (an action that, without further contextual input from the user, may produce erroneous results), the intelligent Agent prompts the user to conduct "housekeeping' ' on selected goals, thereby further removing products returned to the user.
• the system ' s ontological schema is structured specifically to facilitate the conceptualization of decisioning, with the ontology ' s two primary constructs being Classes (with Super Class, Class & Sub-class subsumption) and Attributes (with Attribute and Sub-attribute subswnption); these constructs are "hinged" by intention, which forms the core of human decisioning.
• Such a schema provides less cognitive challenges for content authors utilizing the ontology, thereby reducing authoring errors that are common with complex ontologies, yet providin the power .of linking concepts (for instance, goals to products) across specified domains.
• goals are equivalent to queries, while products become targeted objects of those queries.
• the ontological structure for products mirrors the ontological structure for criteria, or goals; however, they differ regarding the minimum and maximum number of identifiers used to classify each.
• the restriction placed .on goals mathematically forces goals to be a potential subset of products, such that for the universe of goals f Gi. ; ) and products (Pu).
• G K goals which are subsets of products (3 ⁇ 4), so that G ci P K .
• G K goals which are subsets of products (3 ⁇ 4), so that G ci P K .
• a goal will always have I intention and 1 Attribute, but may have 0 of the remaining identifiers; a product will always have at least 1 of every type of identifier.
• subordinate relationships i) Super Class is sub-ordinate to nothing
• vti Sub-attribute is sub-ordinate to Attribute [0093]
• This schema generates output meaningful io humans by enabling machines, via subset calculation, to link richly-defined products to singly-defined goals.
• methods by which subsets are extracted include linear algebra methods, such as multiplication of binary matrices.
• goals need to be purposefully exact— ersus being compound or linguistically complex. It is within the combination of .singly-defined goals that complexity lies, but combination of concepts is handled by the machine, not the user, therefore reducing human cognitive load and increasing the machine's ability to isolate human intention, contradiction and validation.
• Fig. 1 2 visualizes sample product N (13 ⁇ 4 ⁇ ; along with sample goals A, B and C, shown as G A> 3 ⁇ 4 and QQ.
• IxN matrices allow a mathematical result of true when specific instances of products and goals (1), result in [P ⁇ ]*[Gi]-[G f ].
• Matrix multiplication shows that the ontological encoding of GA C . P n and Ge c PN; however, the selected intention within Gu (identified in Fig. 12 with an ⁇ prevents if. from being a subset of PN, such that G & t P .
• the system ' s Content Management System automatically activates an ordinate whenever a content, author selects a suh-ordinate; such auto-classification ensures, that meaningful subsets are not dismissed (a risk that could occur if content authors were responsible for manually tagging subsumption).
• a content author ma select the Subclass Term Insurance when classifying a goal or a product; in this case, the system's Content Management System automatically activates the Class Life and Super Class Insurance, if a content author selects the Sub-attribute Tax-free Growth, the system's Content Management. System automatically activates the Attribute Tax-related.
• the system's ontological schema facilitates behavior mining; specifically, goals and intentions (where intentions are synonymous with Want and Need categories, and are reflected 1 : 1 within the ontology as Intent) intrinsic to user experience are codified with every click, eliminating commonplace predictive modeling that is based on disparate, non-contextual clicks.
• the user is stating behavioral intent, the tool is aggregating & cross-validating those stated intentions; in such a model, semantics enjoys a one-io-one relationship with user action. Pitfalls of prediction are removed.
• Data analytics are not based on. browsing behavior, but rather on state intent as an explicit qualifier to selected criteria or implicit clicks.
• the system ' s ontology represents concepts such that machine-based methods can produce meaningful conclusions specifically to facilitate human decisioning in compressed, timeframes and with- more robust connections, than is otherwise impossible without inordinate human effort.
• the system's ontology is architected to span domains without violating the underlying raeta- model or schema.—hai is, modeling of domains within the system follows a standardized schema versus requiring disparate ontological structures per domain. Genesereth and Nilsson's 1987 definition of domain conceptualization postulates that for domain D, a set of relations R holds unique meaning, and is represented as ⁇ D, R . Modeling within Genesereth and NUsson, disparate domains leverage varied taxonomic. non-taxonomic, partonomic and instantiated relationships; effectively, domain drives structure. Within the present system, however, this one- to-one hypothesis is challenged, such thai structurally:
• the system ' s Content Management System illustrated in 1 108 in Fig. 1 and as indicated by Fig. 13, satisfies three (3) roles: Administrator, Editor and Author. However, the number of Use Cases -is five (5), as the CMS supports two types of Administrators and two types of Authors.
• One person may fulfill multiple roles: for example, within business-to-eonsumer (B2C) domains, one person may serve as Administrator, Editor and Author. Likewise, within business-to-business (B2B) domains, a corporate customer may collapse the. roles of Administrator and Editor. Relationships between roles shown in .Fig. .13 are described as
• This use case contemplates an overall administrator for the CMS tool, spanning all domains, both B2B & 82 C, and having authority over all CMS users.
• the Overall Administrator will be abie to add or remove users of the CMS, and to perform all functions available within the CMS.
• the primary function of the Editor is to update the status of content from pending to eithe approved or rejected.
• the Editor serves in a traditional editing role, and has authority to syndicate authored content to qualified reviewer (Subject Matter Experts) such as functional supervisors, legal professionals, marketing personnel and members of R&D.
• a Domain Editor may designate Subject Matter Experts edit the content directly within the CMS (thereby avoiding the need to copy/paste content into third-party communication tools), by having the Administrator add the Subject Matter Experts, and grant them Editor status.
• Full CMS Authors add content; likewise, they mark existing content as updated or slated for deletion. Since these authors do not directly post updated or deleted content, the production files are secure from erroneous/accidental edits, ' Full CMS Authors have permission to modify content that they did not originally author, since quality control (editor approval) ensures validity of modifications.
• Figs. 14 and. 15 (00981 Still referring to the system's Content Management System, within Figs. 14 and. 15 are shown logic flows. The data inputs, processes, decision -points and outputs within Figs, 14 and 15 are represented with fully contained descriptions, such that persons reasonably versed in reading logic flows may recreate the system's CMS functionality.
• Complementary Goals are used to focus a user on repeating , themes- within their stated goals, and therefore brin focus as they consider results.
• Goal relationships specified within Unchecked Goals mirror those used to- develop Complementar Goal Statements, since the sole difference is that in the former instance, the user has- tailed to check a related .goal, whereas in the later, the Intelligent. Agent is commenting on the relationship of the selected goals.
• contradictory auto statements are limited to discussing 2 related goals at a time, variations of which are displayed in Fig. 11.
• auto-connecting users based cm the percentage similarity of criteria they choose for a decision aid, facilitates collective intelligence; the method identifies users whom a user may not know, but who are. decisioning within the same domain.
• algorithms are used to mathematically assign integer values to user selected criteria and level of importance, as integer attribution can be used to compare decision aids of multiple users and, more important, to flag those users who have decision aids with similarities (based on thresholds which may be administratively raised or lowered over time).
• Fig, 18 there is shown the output table of an algorithm that can be used to- mathematically -assign integer values to users' preferences of a decision -aid, resulting in the ability to compare decision aids and alert users of others who have decision aids with similarities in user preference.
• Need 1 N
• .i for example, for each user of a particular decision aid, has been assigned integer values correspondin to stated importance (as shown beneath the Value header, where 1 and 2 represent Less Important, and 3 and 4 represent More Important), as well as across underlying selection criteria, where 1 indicates user selection of that criteria, and 0 indicates disinterest of that criteria.
• These values are then compared to produce 2803, 2804 and 2805, which represent the similarity percentage per Want or Need, as well as the overall similarity between the decision aids of two users.
• This method .results in a "Connect Me” function, whereb Social.
• Figs. 19 through 22 the formulas represented in Boolean terms are not intended to be specific to any particular programming syntax; not are they presented as limitations, of the invention. Rather, these formulas, are presented as embodiments of calculations from which the system derives user relevancy and intelligent Agent feedback, As example, in both Single-Product Domains and Multi-Product Domains, the states Not. important and More important are used within Intelligent Agent feedback to coach the user toward selecting more criteria—when a majority of goals remain in the default state of Not Important— or to lessen the level of importance in cases where a majority of goals has been designated by the user as More Important. In these instances, example Intelligent Agent feedback reads as follows:
• Equations used to determine if either of the above conditions is true are represented in Fig. 20.
• Figs, 21 and 22 the logic represented in Fig. 21 is referenced within Fig. 25; for Fig. 22, for each Single-Product Domain within the system, a 1 : 1 mapping aligns user-selected criteria with offerings of providers. Although exceptions violate this rule, all data, that is used for comparative purposes is a 1: 1 relationship. Offerings of each provider are entered via a Provider inpu Form— secure, web-accessed tool used by provider personnel (authors) to turn offerings on or off. Although the system's Conten Management System captures this Provider Input Form data offline, the mapping of user-selected criteria to offerings occurs runtime.
• an exception occurs for the fourth Need category within certain domains—called Future Needs, in these instances, providers are not evaluated against the Future Needs category; rather, goals in this category are solely used, for coaching/pedagogical purposes.
• Goal-to-Goal mapping method goals within Future Needs that, are complementary or contradictory to other goals are displayed for content authors to review, append and/or customize.
• Intelligent Agent feedback is displayed, forewarning of contradiction, or emphasizing benefits of similarities.
• each provider within Single-Produci Domains is a standalone set against which a user ' s selected goals are compared, via the percentage-based assessments and the devaluing algorithm displayed in Fig. 22, These comparisons are: executed runtime, coinciding with user input.
• results within the system are returned based on percentage applicability— calculated per category initially, then devalued per category prior to determining an overall percentage applicability (which in itself my be devalued further)— in cases where two results earn equal percentages, then they are displayed in order of assessment. Note that self-learning filters can affect this order, whereby if users access one- result more frequently, then its display may be promoted.
• Algorithms withi the system utilize these constructs in differing -combinations to establish "context ' ' as well as a coach's "fr me-of-mind" (even if. from a Cognitive Load Theory point of view, a user is unaware of intricacies or repercussions inherent within the context).
• Via feedback provided by the Intelligent Agent results "feel right” from a human-coaching point-of- view.
• the algorithm displayed in Figure 22 devalues providers within Single- Product Domains according to rules thai a real-world coach would apply.
• Figs. 23 shows the user action of adding or updating a decision aid to the sy stem:, whereb 10 indicates that the user reviews a list of decision aids, each representing a particular subject, matter and grouped within neighborhoods of content, i further detail, .1 1 indicates that the user provides domain-specific ' filtering criteria, such as search location,, via checkbox, radio button, data entry field, drop down control or other input device common to user-interaction on the web.
• domain-specific ' filtering criteria such as search location,, via checkbox, radio button, data entry field, drop down control or other input device common to user-interaction on the web.
• Block 17 indicates the user's option to save to a member account or to cancel choices associated with the decision aid; and 18 indicates a decision point for the user, allowing further self-representation via subject matters for which additional decision aids are available through the system.
• Fig, 24 there is shown the user action of viewing system results based on user preferences defined in Fig, 23. in particular where 20 indicates the user action of choosing a -decision aid from the se of 1 to m decision aids, where m is limited only by the instances of decision aids the user has added to their member account, since multiples of any single aid may exist within a member's account, effectively resulting in countless instances of decision aids.
• the user may consider intelligent feedback, eCommerce products ' and services, as well as collective intelligence provided by family and friends, the later only in those instances where the user has designated others with whom to share the decision aid, versus marking it private.
• Fig. Still referring to Fig.
• test 21 tests whether the decision aid has been marked social or private, resulting in 22 if social and 25 if private. ' Within the display of a social postings in 22, the test represented by 23 determines if users neither-specified-nor- potentially- known to the user have demonstrated threshold-based similarity in user preferences for the -same decision aid; when true, 24 is displayed, offering to connect the user to similarly- minded, users.
• step 25 tests the domain-type of the decision aid, displaying Single-Product results in 26, and Multi-Product results in 27.
• step 25 tests the domain-type of the decision aid, displaying Single-Product results in 26, and Multi-Product results in 27.
• step 28 displays the display of Market Links, which follow the display of 26 but not 27, since Multi-Product domains specifically do not contain Market Links.
• 29 displays a feature common to both types of domains: Intelligent Agent feedback provides coaching tailored to user preferences and to products deemed relevant to the user, and presented for user consideration.
• Step 30 indicates a user action to view a decision aid, Multi-Product Domains will always fail the test shown in 31,. and will thereby result in 32, which is the display of Multi-Product Domain containers; otherwise, the test indicated by 33 queries a proprietar database for the presence .of -Single-Product Pro viders available to be displayed for user consideration.
• the subset of providers within 33 excludes those providers failing, to meet geo-centric and other domain-specific filters provided by the user; in addition, the subset excludes content that Persona-type and machine-learning filters, as defined below, deem irrelevant to the user; all other providers are contained within the subset -of 33, and therefore must be assessed for degree of relevanc to the user.
• tests 34 and 36 determine whether -the provider via percentage similarity calculations applied against use preferences, an instance of which is described in Fig. 21 , as well as devaluing methods, an. instance of which is described in Fig. 22 is deemed Most Relevant or
• step 40 indicates the user action of choosing to acces either social postings— whereby famil and friends have been granted permission to review and
• a user establishes user . profile data, thereby granting membership benefits to the user, such benefits including user ability to save decision aids for later access, to invite family and friends to review and comment upon certain decision aids, and to keep other decision aids private.
• 51 tests if the user is signed- ⁇ as a member and, if true, captures profile updates, in 55 and displays those updates in 56 for user confirmation.
• 52 requests that the user establish membership.
• the test shown in 53 validates the membership attempt, and if true, captures and displays remaining membership information in.55 and 56; otherwise, if false, 54 warns the users that decision aids can neither be saved nor shared without first establishing valid membership.
• Fig. 28 there is shown the user action of accessing a decision aid defined by another user, and for which permission to review & comment has been granted.
• 60 indicates that a user has selected-— from the View Others " Decision Aids container identified in Fig. 2A, item 1203— to access another member's decision aid.
• 61 indicates the visual display of that decision aid's Wants and Needs, as well as its results, while 62 indicates that a Postings thread is displayed in the feedback container, rather than the Artificial Intelligent Agent's blog that is shown in Figure 4A. item 1401.
• Fig. 28, 65 and 66 respond to the user request to review Intelligent Agent feedback.
• Fig, 29 there is shown the knowledge engineering process to create and map a Provider Input Form based on a decision aid, where the Provider Input Form enables providers a method for posting products and services.
• Such product and sendees may, through similarity matching algorithms for Single-Product Domains, or through ontologica) subset analyses for Multi-Product Domains, be deemed as relevant, to user preferences within the decision aid. Integral to the function of Fig.
• 29 is a knowledge-mapped target domain, constructed via a knowledge acquisition process and populated via electronic forms or surveys,, stored in proprietary database form, and having records within its search domain that reflect, either directly or through related ' products and services, the universe of potential -user-specific criteria tied to Wants and Needs, in further detail, still referring to Fig.
• 70 indicates the content engineer's decision of initiating a new Content input Form based upon one of two proprietary templates; 71 and 72 indicate the content engineer's action of mapping a decision aid's four Want categories, four Need categories, and relevant underlying criteria into the Content Input Form for Single-Product Providers, who are to be evaluated as Most Relevant, Relevant or Of Interest based on user preference; 73 indicates the display of the newly defined Content Input Form for Single-Product Providers. Items 74 through.
• Multi-Product Provider 76 represent the mapping and display of a Content Input Form for Multi-Product Providers, whose offerings of 1 :M products and services are to be evaluated against user preference, and whose offerings may be returned to the user either as Most Relevant or Relevant, since no Of Interest designation is associated with Multi-Product Providers.
• Intelligent Agent feedback triggers based on potentially relevant yet unchecked selections, as well as on the presence of contradictory or complimentary selections within a decision aid, across decision aids, o from a Persona-type baseline, amassed by selections of other users demonstrating similarity in demographics and user preference, in further detail, still referring to Fig.
• step 80 tests if selected criteria are to be assessed within or across decision aids; 81, 82, 86 and 87 determines the appropriate cross-decision aid or cross-user analyses; 83 through 86 and 88 determine the appropriate intra-user and intra-decision aid analyses; in each instance, the extracted Intelligent Agent feedback is displayed in 89.
• the method represented in. Fig. 30 provides occurrences of semantic intelligence reflective of the user's selection criteria and stated importance; such feedback is delivered to the user with the purpose of refining potential results and assisting in better decision-making.
• Fig. 31 there is shown the -method for visually indicating thai a decision aid has been updated.
• the label for a decision aid is visually distinguished whenever intelligent Agent feedback, social posting or ⁇ Commerce- activity has occurred since the user last viewed the decision aid— hether that a d is the users, or that of a family ' member; or friend.
• 90 indicates the user action of opening a decision aid container— either View My Decision Aids or View Others' Decision Aids.
• system processes and queries to determine which decision aids have received updates, and. to graphically highlight all in which the update test proves true.
• Fig. 32 there is shown the method for connecting two potentially unknown users, based on system-identification similarity for a decision aid that both users have included within their member accounts.
• 1 10, 120, 130 and 190 indicate an automated method for assignin integer values to user preferences within a decision aid
• 140 .and 150 indicate a system method and query to compare the values of two aids and determine if conditions are met such thai the similarities are greater- than or equal-to an administrator-assigned threshold
• 160 indicates that each of the two users will be given the opportunity to accept viewing the other's decision aid
• 1 70. 180 represent a system query and method for cycling all decision aids through the comparison routine.
• FIG. 33-a through 33-f example user interfaces are shown to convey a user's progress through major capabilities of . the system.
• the renderings shown in Fig. 33-a and 33-f . are common across Single-Product and Multi-Product Domains; those shown in Fig. 33-b through 33-e represent an embodiment of the Single-Product Domain, highlighting potential user actions.
• a user selects among a list of domains, as shown in 3301 , This walkthrough pursues a user- selected domain of "Explore Senior Living Facilities.”
• the user is shown to have established. a user profile; as such, the user is later presented the option of saving the decision aid before exiting the "View Product Details" interface shown in Fig, 33 -c through Fig. 33-e.
• the user has future access to the decision aid via the '"Manage M Galaxy" container, .shown in its closed state in 3303 of Fig. 33-a, and shown in its open state in 3316 of Fig. 33-f.
• Fig, 33-b the interface for selecting criteria and level of importance for "Explore Senior Living Facilities' * is displayed to the. user.
• filtering information such as location in 3.304, then selecting family members and friends with whom the user wishes to share the decision aid in 3305, the user reviews criteria contained beneath four Want categories and four Need categories, shown in 3306 and .3307.
• the lour Wants categories arc Setting, Living Space, . Lifestyle and Well Being, while the four Needs categories are Care, Medical, Financial and Future.
• each decision aid contains exactly four Wants and four Needs, the number of criteria within each category depends on the topic, and may number from few to dozens or more. Criteria are authored in the form Of attributes or goals, examples of which are shown below for "Explore Senior Living. Facilities";
• the user may change the criteria default stale of "Not Important” to either "Less important” or “More important,” or may choose to leave the criteria as si Not important,"
• the bullsey graphic shown- in 33-07 automatically updates to display, in aggregate, the level of importance for each Want and Need category; in addition, levels of importance drive relevancy of results returned to the user, as described in relation to Fig. 22 and elsewhere.
• the user may view intelligent Agent feedback and results relevant to the selections by clicking 3308 "Vie Results,' ' f 00128] Now referring to Fig.
• the user is presented with Intelligent Agent feedback in 3309.
• Feedback is triggered by user-specified criteria, and includes observations regarding criteria chosen of similar-nature, criteria chosen of eontradictory-nature, and unchecked criteria that may be of value based on other user selections; additional feedback may include suggestions triggered by actions of similarly-minded users, as determined through the adaptive Persona Type method described with Fig. 7.
• the user may click ⁇ 3 ⁇ 4 — Save" to store a piece of Intelligent Agent feedback within the member ' s profile, enabling quick access to the. feedback, as shown in items 3316 and 3317 of Fig. 33-f. Still referring, to the embodiment of the invention of Fig.
• 331 1 shows results returned to the user, grouped within categories of "Most Relevant,” “Relevant,” and "Of interest.”
• the user is shown a detailed description of the result in Fig, 33-d. item 3312.
• the user may choose to print the detailed result or, as shown in 3313, either i) save the result to the member's profile, or ii) share the result with designated family members and friends via the Postings container, shown in Fig. 33 ⁇ e, item 3314.
• the user may return to the main interface, which lists decision aids chosen and available to the user, by clicking the ⁇ Sho Galaxy' * co in 3315.

Landscapes

• Business, Economics & Management (AREA)
• Strategic Management (AREA)
• Engineering & Computer Science (AREA)
• Accounting & Taxation (AREA)
• Development Economics (AREA)
• Finance (AREA)
• Entrepreneurship & Innovation (AREA)
• Game Theory and Decision Science (AREA)
• Data Mining & Analysis (AREA)
• Economics (AREA)
• Marketing (AREA)
• Physics & Mathematics (AREA)
• General Business, Economics & Management (AREA)
• General Physics & Mathematics (AREA)
• Theoretical Computer Science (AREA)
• Management, Administration, Business Operations System, And Electronic Commerce (AREA)
• Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Applications Claiming Priority (2)

Publications (2)

Family

ID=47996515

Family Applications (1)

Country Status (3)

Families Citing this family (7)

Family Cites Families (7)

• 2012
  • 2012-10-01 WO PCT/US2012/058328 patent/WO2013049829A1/en active Application Filing
  • 2012-10-01 EP EP12835351.3A patent/EP2761556A4/de not_active Withdrawn
  • 2012-10-01 CA CA2850606A patent/CA2850606A1/en not_active Abandoned

Non-Patent Citations (2)

Also Published As

Similar Documents

Legal Events