Improvement of interaction between system and user related to advice

Abstract

To provide a computer mounting method for providing a user with notification that can execute system generation on the basis of an interaction between the user and a system, and a computer system.SOLUTION: A method includes the steps of: urging a user to provide a value about user preferences; creating a distribution proposal by distributing the value to a physical entity and a virtual entity; and performing the distribution proposal for allocating a specific value to a plurality of physical entities and an optional virtual entity. At least one of the user preferences defines user evaluation for each of information sources, and the distribution proposal is compared with a comparison configuration including the physical entities and the virtual entity.SELECTED DRAWING: None

Description

[Technical field]

[0001] The present invention relates to the technical field of computer network processing techniques on large data streams. whereby specific user preferences are received and transmitted to multiple physical entities. The allocation proposals between the virtual entities and the data streams, user preferences, and other factors are It is created based on the data. [Background technology]

[0002] Applicant may allocate a finite amount of multiple physical and / or virtual entities. The allocation was based on multiple sources. information and preferences expressed by the user (which may include, inter alia, the evaluation of the source or the The data includes the user ratings and a small amount of geographic and application information for each of the entities. The proposed allocation takes into account user preferences and factors that define and providing the comparison to a user based on the comparison. .

[0003] U.S. Patent No. 1,3935,198 considers the possibility of affecting the vent. Considering the impact of the COVID-19 pandemic on the supply chain (logistics and / or finance), and A system that dynamically manages the supply chain by analyzing external information sources such as supply channels. However, this is not a rigorous system that is entirely focused on the supply chain. The system can be quickly modified to meet other needs or specific user preferences. is not possible.

[0004] Other similar systems, such as those described in U.S. Pat. No. 7,680,719, As is known in the art, Users without the ability to provide personalized preferences or the information sources used Not being properly screened or being unable to access the site due to geographical or other reasons. There is no "practical" filter to filter out unlikely suggestions, There is a lack of user dialogue.

[0005] The present invention aims to address at least some of the problems mentioned above. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] U.S. Pat. No. 1,3935,198 [Patent Document 2] U.S. Patent No. 7,680,719 Summary of the Invention

[0007] In a first aspect, the present invention provides a method according to claim 1.

[0008] In a second aspect, the present invention provides a system as claimed in claim 14.

[0009] In a further aspect, the present invention provides a use according to claim 15.

[0010] Further preferred embodiments and their advantages are described in the detailed description and claims. is. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] As used herein, the following terms have the following meanings.

[0012] As used herein, "a" or "an" refers to both the singular and the plural, unless the context clearly dictates otherwise. For example, "a compartment" refers to one or more compartments.

[0013] As used herein, "including," "comprising," and "having" are synonymous, and are intended to mean, for example, Specifying the presence of the following element and using the Exclude or preclude the presence of additional, non-recited components, features, ingredients, elements, or steps that are disclosed. "Excluding" is a generic or open-ended term that means "not to exclude."

[0014] The recitation of numerical ranges by endpoints includes all numerical values ​​subsumed within that range as well as the recited endpoints. Includes all numbers and fractions.

[0015] The term information source can be file-based, update-based, sync-based, and / or message-based. Any data channel, data service, or other data provider that provides content in a data-based format. It can relate to a stream, a data set, or a data subscription service, thereby Therefore, each source that does not provide data in a pure message is considered to be equivalent to a message base. The updated portion of the data can then be added to one or more new memory locations. The message can be related to a request for a message from a storage device. The information may be provided on request, or via, for example, email, push message, Message notifications, in-app messaging, dedicated web interface, web dashboard The news items may be sent to the user in real time, either as news items or in a feed. Messages are sent in the context of your subscription and / or account with the source provider. It can be delivered within.

[0016] Unless otherwise defined, all terms, including technical and scientific terms, used in disclosing the present invention are intended to be understood as including the following general principles: All terms used herein are as commonly understood by one of ordinary skill in the art to which this invention pertains. Additionally, definitions of terms are included to better understand the teachings of the present invention. .

[0017] The present invention in a first aspect provides a system that provides a user with a response based on the user's interaction with the system. Computer-Implemented Method for Providing System-Generated, Actionable Notifications Using the Method and comprising the steps of: a. Prompt the user to provide multiple user preference values ​​to the system. b. Multiple messages originating from multiple sources are repeatedly received by the server, and multiple A message consists of one or more predefined physical entities and one or more predefined directly and / or indirectly related to the virtual entities defined for The received message is stored as context data. c. At least one factor from each of at least two sets of factors is associated with a physical entity. to each of the virtual entities and each of the virtual entities, thereby Two sets of factors relate to the geographic assessment of physical and virtual entities: The second set of at least two factors is a set of factors between the physical entity and the virtual entity. Relevant to the field of application. d. Providing allocations by allocating values ​​to multiple physical and virtual entities A proposal is created whereby the aggregate share value of the physical entity is non-zero, thereby The aggregate distribution value of the physical entity and the virtual entity is equal to the user-specified value. The allocation proposal includes physical and virtual entities with non-zero allocation values ​​and associations. and each of the non-zero allocated values. e. Providing actionable notices to users, whereby the actionable notices include allocation suggestions; nothing. At least one of the user preferences defines a user rating of each information source. , Thereby, the allocation value is calculated based on the user preferences, the values ​​of the factors and the messages of the information sources. whereby the allocation proposal is based on a plurality of assumed physical entities having non-zero allocation values. The comparison structure includes the concept entity, and the concept entity is compared with the concept entity, thereby A non-zero aggregate share value is equal to a user-predetermined value, whereby the share value of the comparison configuration is equal to the user-selected value. Based on the values ​​of preferences and factors, the actionable notifications are based on the allocation proposals and comparison configurations. is provided based on a comparison with

[0018] The above defined method provides an advantageous solution for entity management in many fields. Given the diversity of ways in which they may be defined, an entity, by definition, Various items such as groceries, cars, raw materials, general supplies, coupons, art, possessions, livestock, etc. From tangible tradable instruments to stocks, common stocks, bonds, cash, real estate, valuables, and investment funds These can vary widely, ranging from more cumbersome tradable commodities such as treasury bills, etc. The distribution of values ​​of such entities may include, for example, the location of storage devices, the owner, the price, the You can assign costs for the installation, characteristics of the storage, transmission of the entity, etc. The optimized allocation of such parameters / values, some explicit, others implicit or indirect In general, it is highly dependent on many agents. The first agent can be user preferences, which often influence the allocation. It has a strong deterministic property and serves as a guide for other agents to carry out optimization. In particular, the user's preferences in this case should include the use of at least some other agent. It can act in a reflective manner. The second agent is the multiple information sources, as already mentioned, among others the user-selected The preference defines user ratings of information sources that guide the allocation process, and the ratings are the ratings that each information source receives in the allocation process. A user may prefer certain sources over others. (e.g. Twitter feeds the official Dow Jones news) There are sub-ratings that further specify the weighting given to particular subsets of information from some sources. It is possible to exist. Third, the agent can define a set of factors as "factors" (elements). Factors can be assigned to each entity and are used to measure the specific characteristics of the entity. At least one of the factors defines the geographic information of the entity, and at least At least one factor defines the application area of ​​an entity. Geographic information is typically is the country, region, or continent representing the entity (but not the city, state, or (It can be one or more of several groups of materials. For example, raw materials For physical objects, the place of origin (area of ​​manufacture or the like) or actual physical location In the case of a virtual object, it may be the location of its owner, or the location of the associated The location of a physical entity (e.g., in the case of stocks, the location or territory of the company issuing the stock) ) can be expressed in terms of the application field. As defined above, stocks, raw materials, property, food, medical equipment, etc.) or more generally, Such factors can reflect the technical field to which the company belongs, such as customer care, technology, etc. It is known that this will often have a strong impact on optimal allocation. For example, food In the case of the allocation of entities, the geographic location is determined by the “category” ( meat, dairy products, grains, preserves, etc.) would be important. The same is true for e.g. The same applies to the construction of investment portfolios, where location affects the security of physical goods. and thus affecting their reliability / value or, in the case of stocks, volatility. Such factors may affect the entity (physical It is important that the objective (whether objective or virtual) is defined. In addition, the entity's factors determine how the computer system affects the entity. It defines a basis that allows for the establishment of associations with specific messages that can be This simplifies the calculation of the property share. In these cases, there is limited available cargo space (weight and / or volume), limited Whether this is due to refrigeration, manpower, and / or even budgetary constraints Regardless, user-defined constraints must often be imposed on the distribution values. Thus , which are set as the maximum of the aggregate distribution values ​​for the entity. Finally, the system creates a list of all entities with non-zero allocation values ​​and the Generates an allocation proposal that essentially includes the associated allocation values ​​of the properties. This can be changed or with or without correction (e.g., information on the person not yet in the system) (if you know of a particular shortcoming or simply have a particular weakness due to previous experience) ) and authorized persons for further analysis. However, before the allocation proposal is provided to users, one or more These “known” comparative structures are used to compare the proposed allocations to those that have been used in the past. Configurations generated by the old system, user-generated plans, and external information The requirements provided to the user can be obtained from the so-called benchmark allocation. To meet the requirements, the allocation proposal must "outperform" the comparative structure based on certain criteria. do not have. Finally, the allocation proposals are provided as actionable notices and have practical (technical) effects. understand that the It can be understood.

[0019] As used herein, the term "value" or "allocation value" refers to specific information about an entity. These values ​​can be understood to represent the The amount of the entity that is proposed to be "used" may be the amount of the entity that is proposed to be "used" in the proposed manner.

[0020] All of the prior art systems require a user-to-system interaction to generate allocation proposals. For example, systems generally do not allow for such complex interactions between users and However, human experience is such that certain information may be subjective. It must be understood that the information sources can be properly evaluated to guide the system. For example, when distributing emergency supplies to deal with a humanitarian crisis, certain trusted accounts The company has been trying to make the same kind of payment system available to its users, but it is not able to provide urgent content as quickly as newspapers. By providing more bias than other sources, such as classic sources, users can This could provide great value on "fast" social media sources such as Conversely, more volatile news sources (Twitter, Facebook, etc.) or self-published web content) is a typical self-publishing if time is not an issue. , rated lower than traditional news sources, or other more regulated channels. It is possible.

[0021] In a preferred embodiment, the user preferences include at least a regularization type preference and a regularization amount preference. . Preferably, the regularization type preference is a regularization without variable selection, or a regularization without variable selection. Related to any regularization with selection, More preferably, the regularization without variable selection is Tikhonov regularization or ridge regularization. Related: More preferably, the regularization involving variable selection is related to Lasso regularization. Further, preferably, the regularization amount preference is a value indicating the strength of regularization, more preferably a real value, Related to.

[0022] Here, "regularization" refers to the reduction in the number of sources of information that provide messages that affect the evaluation output. A mechanism for evaluating a source's message, taking into account its reliability and / or importance. This mechanism works by detecting messages with a high level of "noise", i.e., high uncertainty. Although messages are constantly being generated, the majority of messages are generated by evaluation. contributes little, if any, to the final value of the Thus, regularization is based on a stable mathematical concept and is generated by evaluation. This reflects the level of distrust regarding the output being sent. Regularization can therefore be understood as a three-step approach: Some notion of normality for each source, typically the absence of a message, or Define "zero signal." The stream of messages from all sources is the sum of content and noise, or Model it as "real signal + noise." To get closer to normal, we filter out noise by delivering the output of the evaluation, or "whole signal." Remove.

[0023] In one variation, the regularization involves preserving all messages from all sources while The incoming arrivals are then evaluated according to the uncertainty, preferably updating the evaluation criteria for each arrival. The impact of the message is automatically adjusted to reduce it. Next, the trigger for the notification is any information The regularization can be done by applying the This allows the user to preferably select normalization amount preferences. The strength of normalization can be controlled via the corresponding and / or related preferred In one embodiment, the regularization type preference is related to regularization without variable selection, and regularization without variable selection is is preferably associated with Tikhonov or ridge regularization, and the amount of regularization preference is The intensity of the signal is related to a value, preferably a real value.

[0024] In another variation, regularization involves adjusting messages according to uncertainty while adjusting other messages according to uncertainty. The message is deemed too uncertain or too insignificant. This corresponds to automatically suppressing any information source that is used for a certain period of time. Allocation proposals can be based only on messages from sources that were not suppressed, and This allows the regularization to further moderate the influence of unsuppressed messages. This allows the user to preferably control the strength of regularization via the amount of regularization preference. In a corresponding and / or related preferred embodiment, the regularization type preference is Related to the regularization using variable selection, the regularization using variable selection is preferably the Lasso regularization. Relatedly, the regularization amount preference relates to a value, preferably a real value, that indicates the strength of the regularization.

[0025] Regularization provides users with an additional useful tool for evaluating the value of various information sources. , which greatly aids the task of allocating values ​​to entities.

[0026] In a preferred embodiment, the value allocated to the entity of the allocation proposal is equal to at least one user. The optimization model is further calculated based on the optimization model selected by the user, the optimization model being at least A list including the Black-Litterman model, the Markowitz model, and the minimum variance model. is selected from.

[0027] In a preferred embodiment, the information sources are direct context information sources and structured surrounding information sources. It includes information sources, unstructured peripheral information sources, and inferential direct context information sources. Direct contextual source data is enriched through feature engineering Preferably, this is achieved by utilizing a priori direct context information sources and / or configurations. Built-in peripheral source data is enhanced through feature engineering.

[0028] In a preferred embodiment, the plurality of information sources includes at least one unstructured peripheral information Including the source, At least one message originating from an unstructured peripheral source is raw text. Contains the string, Raw text strings to evaluate the relevance of messages with respect to contextual data comparing the message to the contextual data using natural language processing of the , This allows the allocation value to be calculated further based on the evaluation.

[0029] In a further preferred embodiment, messages from unstructured peripheral sources are not calculated. The calculated allocation values ​​are processed before use, whereby the processing is carried out through a self-learning artificial intelligence. It involves structuring messages from unstructured peripheral sources, whereby artificial intelligence It operates under user-tunable hyperparameters.

[0030] In a preferred embodiment, the physical and virtual entities each have a context. Based on at least historical information between physical and virtual entities in the store data The allocation value of the allocation proposal is then assigned a predictability rating based on the predictability rating. The calculation is further based on the

[0031] Knowledge of the "predictability" of a particular entity can help predict its further development. and may be useful in creating a certain balance or security in the overall allocation proposal. This can be done.

[0032] In a preferred embodiment, the prompted user preferences include one or more constraints, which are Physical and / or virtual entities with non-zero allocation values ​​in the allocation proposal User-defined minimum predictability rating of the property and / or non-zero in the allocation proposal User-defined minimum aggregate budget for physical and virtual entities with distribution values Includes at least a measurability rating.

[0033] As mentioned above, ignoring the predictability ratings of entities can lead to very risky allocations. Suggestions may be generated, which may be efficient or advantageous, but all are subject to the plan. Only under the assumption that the future impact of each entity will be moderately affected. Even when a large number of different entities are involved, there is still a large difference in the potential variance of each entity's impact. (For example, catastrophic harvests, production problems, etc. may affect certain entities more than others. This is not reflected in the forecast. Therefore, depending on the situation (urgency, location, etc.) ,Users may be allowed to rate the allocation proposals ,higher or lower predictability. can.

[0034] In a preferred embodiment, the prompted user preferences are used in one or more of the steps to generate allocation proposals. The constraint includes a number of constraints on the number of physical entities and and / or at least the maximum and / or minimum number of virtual entities; and / or The maximum and minimum allocation values ​​of the physical and / or virtual entities in the allocation proposal. and / or minimum value.

[0035] In a preferred embodiment, a user preference is one or more objects that have a non-zero past value. historical allocation mix, including physical and / or virtual entities; This allows us to calculate the rebalancing cost of converting the past allocation structure into the proposed allocation. Based on the context data adapted to the can be.

[0036] Historical allocation structure includes the "current" or "latest" (or expected) allocation structure; Entity (or at least either the proposed allocation and / or past allocation structure) Represents the last known value for the entity that has a non-zero distribution value in Value may vary depending on administrative, governmental or other factors such as taxes, shipping costs, etc. It will result in certain costs in practice, which will not necessarily be "monetary" The cost of time or labor (labor costs) to enact a particular change, and Note that can be the cost of space (volume / weight). Present allocation suggestions These costs may affect the merits of the proposed allocation. ,must be taken into account by the user (e.g., too high time costs will lead to other can negate all of the advantages).

[0037] In a preferred embodiment, the prompted user preferences are used to generate at least one allocation proposal. The constraints are the rebalancing costs for converting the past allocation structure into the allocation proposal. This is the maximum value.

[0038] In a preferred embodiment, the context data is further processed via ensemble learning. Combining multiple learning algorithms improves predictive preferences across treatments. can be.

[0039] In a preferred embodiment, user preferences are calculated based on a small amount of contextual data based on the calculation of a distribution value. At least one constraint, whereby the constraint specifies whether a message containing context data The relationship between the period during which messages are generated and the user evaluation of the source of the message, including the context data. Users may, for example, choose to disable certain features in order to simplify the evaluation process. It is possible to focus on information for a specific period of time.

[0040] In a second aspect, the present invention provides a method for providing a user with a system based on an interaction between the user and the system. In a computer system for generating an actionable notification for A server, the server having a processor, a tangible non-volatile memory, and instructions for the processor. and a program code present in memory for instructing a user device and one or more a server including a connection means for connecting to a remote server; b. A user device, the user device comprising a processor, a tangible non-volatile memory, and Program code that resides in memory to instruct the processor and display information to the user and preferably an input means for receiving user input from a user. a connection means for connecting to a server via a computer network; c. one or more remote servers, each of which is a plurality of information sources; Each of the one or more remote servers is associated with at least one of the a remote server including a connection means for connecting to the server via a data network; d. at least one computer-readable medium, The readable medium is accessible to a server and includes a database, the database being At least one computer readable medium including text data and user preference data. a computer readable medium, preferably contained within a server; Including, The computer system is configured to generate an actionable notification to a user. generating an actionable notification, prompting a user to provide values ​​for a plurality of user preferences to the system; The server repeatedly receives multiple messages originating from multiple sources. a plurality of messages are sent to one or more given physical entities and one or more Messages received that are directly and / or indirectly related to a number of given virtual entities the message is stored as context data; At least one factor from each of at least two sets of factors is applied to a physical entity. a step of assigning to each of the first at least one virtual entity and each of the virtual entities, Two sets of factors relate to the geographic assessment of physical and virtual entities: The second set of at least two factors is the application of the physical entity and the virtual entity. Steps related to the field, Providing allocation suggestions by allocating values ​​across multiple physical and virtual entities creating a physical entity having a non-zero aggregate share value and The aggregate allocation value of the property and the virtual entity is equal to the user-specified value, and the allocation proposal is non-zero. and associated non-zero allocations. each of the values; Providing an actionable notification to the user, the actionable notification including an allocation suggestion. Well, steps and Including, whereby at least one of the user preferences defines a user rating of each information source; The distribution value is calculated based on the user preferences, the values ​​of the factors and the messages of the information sources, and the distribution value is The distribution proposal includes a plurality of physical entities and virtual entities having non-zero distribution values. The comparison structure is compared to the comparison structure, and the non-zero aggregate distribution value of the entity in the comparison structure is equal to the user-predetermined value. Specifically, the allocation value of the comparison configuration is based on the user preferences and the values ​​of the factors.

[0041] In a third aspect, the present invention relates to the use of a method according to the invention in a system according to the invention. This relates to.

[0042] The present invention is further described by the following non-limiting examples which further illustrate the present invention. It is not intended to be, and should not be construed to be, limiting the scope of the present invention. It shouldn't be. [Example] EXAMPLES

[0043] <Physical Entity Collection> In this example, the present invention allows a user to monitor a collection of physical objects. Relatedly, a physical object collection may include multiple objects stored or located in multiple locations. Physical goods or objects, e.g., gold, diamonds Books about Mondo, (vintage) automobiles, wine, whiskey, fine art paintings, musical instruments, or jewelry They can be valuables that have a qualitative value, but they can also be valuables whose value is attached to the owner, such as bearer bonds. It may also relate to valuables related to the rights granted to the The management object may be related to real estate, for example, the object may be a house, building, Can correspond to a property, apartment, apartment block, garage, or garage block. In a preferred exemplary embodiment, each object is stored in a database. In one embodiment, a physical object corresponds to a physical entity record and vice versa. However, there are cases where excessive consumption is seen, such as in the case of automobiles, wine, whiskey, fine art, and musical instruments. Perishable, susceptible to damage due to temperature or high moisture-related incidents Here, the first allocation proposal may relate to a given time period, e.g. Maximize the total number of objects across locations that remain intact over the years Such allocation proposals may involve the use of available funds from different sources. Based on the message originating from one of the sources, taking into account measurements and forecasts of different positions, ,Whenever any location is likely to suffer an accident, a notification can be ,generated. However, the second allocation proposal requires that at least one object is , for example, related to maximizing the chances of remaining intact over a year. This is because, for example, a hard disk drive or flash drive can This may involve the inclusion of highly confidential information that is replicated across multiple locations. This allows the hard disk drive or flash drive to retain the confidentiality required. Depending on the amount of data that can be or will be connected to a computer network, Therefore, the data on the drive can be encrypted or Such highly secret information can be, for example, passwords, authentication information, Such allocation proposals may relate to a single position. A prediction or measurement of the single location at which most of the object is stored. Unless the latter is true, it may lack relevance and therefore does not require notification. Whether this is the case will depend on the fact that indicators of an incident are detected at that single location. This can be illustrated by the strong change in the output of the allocation proposal. In an embodiment, one information source may be a signal such as an alarm or incident detection at an object location. Alarm services that provide location and / or building integrity and safety related alarms, such as Another source of information may be a live text service that provides news messages. This allows the name of the location and / or the name of the building associated with the location to be Previous detection may indicate that the message is relevant in light of the allocation proposal.

[0044] The advantage of the present invention is that it allows users to easily communicate what is and is not relevant. Regularization is about controlling how much knowledge is needed. For example, from a single source about a fire at a location News messages in the,dataset can be considered unimportant and,based on regularization, can be considered outliers. while multiple messages with very similar content may be highly relevant. This allows regular The strength of the filtering allows the user to set a threshold according to their preferences. In addition, the user can set the variable You can indicate whether you prefer regularization without selection or with variable selection. As a user, you can view all messages that are potentially related to a physical object collection. Regularization relieves users from the burden of manually examining the information provided by various sources. The present invention further provides a useful tool for evaluating the value of Which subscriptions to sources (or sources not) should continue and which subscriptions can be removed This is valuable when deciding whether to EXAMPLES

[0045] <Portfolio of tradable products> In this example, the context data is a complex set of data related to at least two physical entities. It refers to all information related to a portfolio that contains a number of tradable instruments. The physical data includes a plurality of data records including at least two physical entity records, each of which includes Each physical entity record is associated with a physical entity. Physical investment products, e.g. gold, diamonds, (vintage) cars, wine, whiskey, Fine art paintings, collector's items, musical instruments, jewelry, bearer bonds, houses, buildings, apartments, It can be a part block, a garage, or a garage block. Beyond the physical entity record Therefore, a portfolio may contain data records relating to entities that are not physical entities. Each data record may or may not contain a structured set of attributes. An attribute is the type of entity, e.g., stock, equities, bonds, cash , real estate, valuables, investment funds, etc. Another attribute is the entity Yet another attribute can indicate the start date of the promise associated with the entity. Or it can relate to an end date. As in example 1, another attribute could be, for example, a street address. or the location of an entity, specified in terms of a set of GPS coordinates. Another attribute is the natural language description of the entity, which can be taken into account through natural language processing. This allows for the description of

[0046] The proposed allocation is related to the technical objectives of the portfolio. The composition of the proposed allocation depends on the It involves training on multiple parameters on history. The allocation proposal is based on the Black-Litterman optimal The information source may or may not be related to the application of the information. It includes pre-structured data, which contains records with inputs, and unstructured data, which contains raw text. Pre-structured data includes unstructured data as well, both at the company and aggregate levels. , which provides indicators of news intensity, sentiment towards news, and news readership. The generated data can be associated with the index. Users can also The sentiment score was developed to respond to developments in the use of the app and in official press conferences. It provides the possibility to change the source of information, where opinions are typically based on some past It is measured in relation to standards, but it places more emphasis on trusting those who shape certain public opinions. The relevant criteria can be modified in this way. The process flow can be, for example, Or it can be modified similarly to occur more frequently.

[0047] Therefore, automated processing, together with the added value of the monitored portfolio, In particular, the present invention provides a comprehensive alternative to purely manual processing. provide partial alternatives to sub-decisions and provide information generated in the process Any significant changes in the forecast are reported to the user, e.g., the portfolio manager. The user can then rebalance the portfolio. The present invention further relates to a method for determining whether a particular entity and / or associated financial instruments are In addition, the user can identify whether a particular entry has been reviewed. You can be alerted to the amount of negative news accumulating around you. For example, if the stocks of some companies suddenly fall, the affected affiliates and sectors will be Based on the target, an indirect relationship can be established involving multiple entities.

[0048] The manner in which notifications can be generated for the user is further similar to Example 1. Also, in this example, the user can select all the potential related Regularization relieves users from the burden of examining messages from various sources. Further provided is a useful tool for assessing value. Should subscriptions to sources (or not) continue and which subscriptions can be removed? It is valuable in determining whether you can EXAMPLES

[0049] <Crisis supply allocation> In this example, contextual data can be used to inform local crises such as famine, disease outbreaks, or natural disasters. It relates to information about a situation. In this case, the source of information is traditional media (television, newspapers, online Line reports), social media (Twitter, Facebook, Instagram ram, Tencent, Weibo, Youtube, etc. The services may relate to food, medicines, infrastructure supplies (tents, etc.), financial support, etc. The user can then provide specific values ​​in response to system prompts. In these cases, unstructured information is The source can quickly identify risk situations at a local level (e.g., a particular village where an outbreak has occurred). While the larger media coverage may not have covered these They may overlook the source or refuse to report it without further verification. Thus, users may give such sources a greater rating. Other information that may be communicated would be geographic details and the nature of the crisis. The system may, for example, monitor the supply of grain, medicines, and other items in a surrounding area (country, region, city, etc.). Medical personnel from an organization (Doctors Without Borders, ...), transport capacity (cargo with capacity Y in country Z) A large amount of entity-based data with entity-based information (e.g., number of planes, ...) The database will be updated through the influx of information from the sources. Please note that the following points will be updated: Based on this, the system allocates resources to deal with the crisis on a time basis. A proposal can be generated that distributes a value across multiple entities. whether they are used (e.g., that a particular supply is provided from a particular location within a certain time period) If it is not possible to do this, these will not be used), and how many of them are used? It can be shown what information is provided and how it is provided.

[0050] The present invention is not limited to the above implementations, but should be construed as being within the scope of the appended claims. It is assumed that modifications can be made to the presented embodiment without the need for modification. While the present invention has been described with reference to the allocation of goods, products, and the like, both physical and virtual, Obviously, other applications are possible.

Claims

1. System-generated, actionable notifications based on interactions between a user and the system.

1. A computer-implemented method for providing a user with a ... computer-implemented method, the method comprising: a. prompting the user to provide values ​​for a plurality of user preferences to the system; Steps and b. A server that receives multiple messages from multiple sources in a recursive manner The plurality of messages are transmitted to one or more predetermined physical entities in a single communication step. directly and / or indirectly associated with one or more predetermined virtual entities, The received message is stored as context data; c. at least one factor of each of at least two sets of factors is applied to the physical entity; a step of assigning a first identity to each of the virtual entities and each of the virtual entities, the step comprising: The at least two sets of factors are the physical entity and the virtual entity. and a second set of at least two factors related to a geographical assessment of the physical entity. and d. By allocating values ​​to a plurality of said physical entities and said virtual entities. generating an allocation proposal based on the aggregate allocation value of the physical entity, and the aggregate distribution value between the physical entity and the virtual entity is user-specified. value, and the allocation proposal includes physical entities and virtual entities that have non-zero allocation values. an entity and each of said associated non-zero distribution values; e. providing an actionable notification to said user, said actionable notification comprising: including the allocation proposal; Including, At least one of the user preferences defines a user rating of each of the information sources; The distribution values ​​are the values ​​of the user preferences, the factors, and the messages of the information sources. and the allocation proposal is calculated based on a plurality of physical entities having non-zero allocation values. and comparing the virtual entity to a comparison configuration including the virtual entity, A non-zero aggregate share value is equal to the user-predetermined value, and the share value of the comparison configuration is equal to the user-predetermined value. Based on the preferences and the values ​​of the factors, the actionable notification is based on the allocation proposal and the ratio. Based on a comparison with a comparative configuration, 23. A computer-implemented method comprising:

2. the user preferences include at least a regularization type preference and a regularization amount preference; The regularization type preference is preferably a regularization without variable selection or a regularization with variable selection. Related to any of the regularization The regularization without variable selection is preferably Tikhonov regularization or ridge regularization. Related to regularization, Said regularization involving variable selection is more preferably related to Lasso regularization, The regularization amount preference is preferably a value, more preferably a real value, indicating the strength of the regularization. , related to, 2. The computer-implemented method of claim 1.

3. The values ​​allocated to the entities of the allocation proposal may be based on at least one user selection. The optimization model is further calculated based on an optimization model of at least Black- Selected from a list including the Litterman model, the Markowitz model, and the minimum variance model.

3. A computer implementation according to claim 1 or claim 2. method.

4. The information sources are structured into direct context information sources and structured surrounding information sources. The direct context information includes a set of peripheral information sources that are not related to the direct context information, and a set of a priori direct context information sources. The source data is enhanced through feature engineering, preferably The data from the direct context sources and / or the structured surrounding sources are Enhanced through characteristic engineering, as claimed in claims 1 to 5.

4. The computer-implemented method according to any one of claims 3 to 3.

5. the plurality of information sources includes at least one unstructured peripheral information source; At least one message originating from the unstructured peripheral information source is raw text. Contains the string, to evaluate the relevance of the message with respect to the context data. Natural language processing of text strings is used to correlate the message with the context data. comparing, The allocation value is calculated further based on the evaluation. A computer implemented method according to any one of claims 1 to 4. 。

6. Messages from the unstructured peripheral information sources are distributed using the calculated distribution values. The unstructured peripheral information is processed beforehand, and the processing is performed by using a self-learning artificial intelligence to The artificial intelligence includes structuring the message of the source, and the artificial intelligence includes a user-tunable hypervisor. The computer according to claim 4 or claim 5, characterized in that it operates under the following parameters: Computer implementation method.

7. The physical entity and the virtual entity each include the context data. based on at least past information of the physical entity and the virtual entity in , the allocation value of the allocation proposal is assigned a predictability rating, The method according to claim 1 to claim 5, further comprising the step of calculating the degree of measurability based on the degree of measurability.

7. The computer-implemented method of claim 6.

8. The prompted user preferences include one or more constraints, the constraints being applied to the allocation proposal. The physical entity and / or the virtual entity having a non-zero distribution value in and / or said constraints include at least a user-defined minimum predictability rating of said properties. The physical entity and the virtual entity having a non-zero allocation value in the allocation proposal are and a user-defined minimum aggregate predictability rating for each of the entities. A computer-implemented method according to any one of claims 1 to 7.

9. The prompted user preferences include one or more constraints for generating the allocation proposal. , the constraints are physical entities and / or entities that have a non-zero allocation value in the allocation proposal. or at least includes a maximum number and / or a minimum number of virtual entities; and / or The constraints may be applied to the physical entities and / or the virtual entities in the allocation proposal. The maximum and / or minimum values ​​of the distribution values ​​of the identity are included.

9. A computer-implemented method according to any preceding claim.

10. The user preferences may include one or more of the physical entities having non-zero past values. and / or a historical allocation structure that includes the virtual entity; calculating a rebalancing cost for converting the past allocation structure into the allocation proposal; Based on the context data, an objective cost function is is defined as 10. A computer implemented method according to any one of claims 1 to 9.

11. The prompted user preferences include at least one constraint for generating the allocation proposal. The constraint is the rebalancing cost for transforming the past allocation structure into the allocation proposal.

11. The computer-implemented method of claim 10, wherein the maximum value of

12. The context data is further processed via ensemble learning.

12. The computer-implemented method of claim 1 .

13. The user preferences define constraints on the context data based on the calculation of the allocation values. and wherein the constraint includes at least one of the message including the context data. the time period during which the message is generated and the source of the message including the context data; and a user rating of the user.

13. The computer-implemented method of claim 12.

14. A computer system for generating an actionable notification to a user, said system The system is based on an interaction between a user and the system, the system comprising: a. a server, said server comprising a processor, a tangible non-volatile memory, program code present on said memory for instructing said user device; a server including connection means for connecting to one or more remote servers; b. said device of said user, said device having a processor and a tangible non-volatile a memory; and program code present on the memory for instructing the processor. a screen for displaying information to the user, and preferably a means for receiving user input from the user; and an input means for connecting to said server via a computer network. and a connection means for connecting the c. the one or more remote servers, each of the remote servers being A remote server associated with at least one of a plurality of information sources, The server includes a connection means for connecting to the server via the computer network. A remote server including: d. at least one computer readable medium, said at least one computer The data-readable medium is accessible to the server and includes a database, The base includes context data and user preference data, and the at least one computer The computer readable medium is preferably included in the server; Including, The computer system is configured to generate the actionable notification to the user. generating the actionable notification comprises: prompting the user to provide values ​​for a plurality of user preferences to the system; Top and The server repeatedly receives multiple messages originating from multiple sources. the plurality of messages are transmitted to one or more predetermined physical entities and or a plurality of predetermined virtual entities, the received message is stored as context data; At least one factor of each of at least two sets of factors is assigned to the physical entity. a step of assigning a first pre-defined property to each of the virtual entities and each of the virtual entities, The at least two sets of factors are related to the physical entity and the virtual entity. The second set of at least two factors relate to the geographical evaluation of the physical entity. and By allocating values ​​to the plurality of physical entities and the plurality of virtual entities, generating a distribution proposal, the distribution proposal being based on the aggregate distribution value of the physical entity being non-zero; The aggregate distribution value between the physical entity and the virtual entity is equal to a user-predetermined value. Preferably, the allocation proposal includes the physical entities and the virtual entities having non-zero allocation values. and each of said associated non-zero distribution values; providing an actionable notification to the user, the actionable notification being including an allocation proposal; Including, At least one of the user preferences defines a user rating of each of the information sources; The distribution value is a function of the user preferences, the factors, and the messages of the information sources. The allocation proposal is calculated based on a plurality of physical entities having non-zero allocation values. and comparing the comparison configuration with a virtual entity and the entity of the comparison configuration. The non-zero aggregate share value is equal to the user-predetermined value, and the share value of the comparison configuration is based on user preferences and the values ​​of the factors; A computer system comprising:

15. The system according to claim 14, further comprising: Use of the methods described.