Abstract: A system and method are disclosed for a low-touch centralized system to predict service level failure in a supply chain using machine learning. Embodiments include receiving only historical supply chain data from an archiving system for one or more supply chain entities storing items at stocking locations, predicting one or more supply chain events during a prediction period by applying a predictive model to a sample of historical supply chain data, calculating an occurrence risk score for at least one of the one or more supply chain events and indicating a possibility that the at least one of the one or more supply chain events will occur, generating one or more alerts identifying at least one item and at least one alert stocking location, rendering an alert heatmap visualization comprising one or more selectable user interface elements, and provide one or more tools for initiating corrective actions to be undertaken.

Abstract: A method of training a neural network to approximate a forecasting error of a passenger-demand forecasting model that includes calculating, using the forecasting model, a historical passenger demand forecast for each key level in a set of key levels and for each departure date in a set of historical departures dates; applying a dropout model to the historical passenger demand forecasts to create a training sample; training, using the historical passenger demand forecasts and the training sample, the neural network, to approximate forecasting errors associated with the forecasting model; calculating, using the forecasting model, a future passenger demand forecast for each key level in the set of key levels and for each departure date in a set of future dates; and approximating, using the trained neural network, the forecasting error associated with the future passenger demand forecasts for the second set of departure dates.

Abstract: A computer-executable method, a computer system and a non-transitory computer-readable medium are provided for causing electronic marketing communications of one or more promotions to be generated on a mobile computing device associated with a consumer. A method includes programmatically retrieving promotion data indicative of a plurality of promotions from a computer memory. The method includes determining, using processing circuitry, a promotion score for each of the plurality of promotions. Each promotion score is determined based on consumer profile data, stored consumer activity data, and at least one of: current consumer activity data, current local context data, or predicted consumer activity data. The method further includes outputting indications configured to generate electronic marketing communications associated with the plurality of promotions based on the promotion scores of the plurality of promotions.

Abstract: Systems and methods for a global marketplace, authentication service, and exchange for asset tagging, weighing, measuring, authenticating, and management. Users connect any item to the platform by scanning a tag affixed to the item. Scanned items are registered to a user's account. The platform deploys Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) in order to promote efficient and effective marketing and monetization of a user's registered items, in addition to providing organizational, repair, and maintenance services for any item tagged by a user.

Abstract: A method is provided for determining a social rank of a node in a social network, the social network including a plurality of nodes connected by relationship chains. The method includes determining a user corresponding to at least one of the plurality of nodes in the social network, determining a connection structure of the relationship chains between the plurality of nodes, and determining, according to the connection structure of the relationship chains between the plurality of nodes, a social rank of the user corresponding to the at least one of the plurality of nodes.

Abstract: Provided is a system, method, and apparatus for automatically preparing documents for a multi-national organization. The method includes generating a document preparation workflow based on multi-national organization data for the multi-national organization, the document preparation workflow including a sequence of nodes based on multi-national organization data and rules data, the multi-national organization data including at least one multi-national organization chart and transactions associated with the multi-national organization, configuring at least one node of the sequence of nodes by identifying at least one role associated with the at least one node of a plurality of roles, and executing the document preparation workflow by automatically performing a plurality of actions associated with the sequence of nodes based on the multi-national organization data and the rules data.

Abstract: A method for integrating a machine learning (ML) model that impacts different factor groups for generating a dynamic recommendation to collectively optimize a parameter is provided. The method includes (i) processing a specification information and operational data associated with a demand management service obtained from client devices (116A-N), (ii) training the ML models with processed specification information and the operational data to obtain a trained ML model that includes an anticipation ML model that optimizes demand parameter or recommendation ML model that generates recommendation for optimizing a factor group, (iii) integrating the trained ML model with the ML models by setting an output of a first ML model as a feature of a second ML model and (iv) determining a demand of a product using the trained ML models and quantifying probabilistic values that signify prediction of the demand.

Abstract: The present teaching relates to method, system, medium, and implementations for machine learning. Upon receiving input data associated with a time series, hidden representations associated with the time series in a feature space are obtained and used to generate a query vector in a query space. Such generated query vector is then used to query relevant historic information related to the time series. The query vector and the relevant historic information are aggregated to generate at least one queried vector, which is aggregated with the hidden representations to generate enriched hidden representations that enhance the expressiveness of the hidden representations.

Abstract: An electronic computing device includes a processing unit and system memory. The system memory includes instructions which, when executed by the processing unit, cause the electronic computing device to receive data associated with one or more customers of an institution. The data is received from one or more other electronic computing devices. The received data is analyzed to identify grammatical elements in the data. Relationships are derived between a plurality of the grammatical elements. At least one derived relationship is used to update a profile for a customer. At least one derived relationship is used to identify a customer for which a remedial action is warranted.

Abstract: A method and system analyzing transactional data are provided. The method includes gathering transaction data related to a first physical entity; extracting a first set of features representing the first entity; gathering transaction data related to a second physical entity, wherein the at least one second physical entity is indirectly related to the first physical entity; extracting at least a second set of features associated with the at least one second physical entity; matching the first set of features to the second set features; and clustering the first set of features and the second set of features, when the first set of features matches the second set of features.

Abstract: The present disclosure provides methods and systems for tracking a shopper route at a retail enterprise. Location information associated with assets can be collected at a retail location, from which a detailed route through the retail location may be recreated and overlaid on map data reflecting a retail location layout. Further analysis may be performed on the route. Additionally, the route may be overlaid on a map, including business context information and point-of-sale transaction information, allowing for various metrics and metric visualizations to be generated that can be further analyzed to achieve various objectives.

Abstract: In one implementation, a method for determining accurate store traffic levels includes receiving, at a computer system, historical store data for a physical store that was generated by internal data sources located within the physical store; generating, by the computer system, a historical store model that correlates values from store data with guest traffic levels within the physical store; receiving, by the computer system, current store data from one or more of the internal data sources; determining, by the computer system, current guest traffic level in the physical store based on the current store data and the historical store model; and outputting, by the computer system, the current guest traffic level in the physical store.

Abstract: Systems and methods of the present disclosure enable improved content filtering by determining a profile state of the user profile based at least in part on at least one attribute of the user profile and determining a user profile criterion associated with the user profile based on the profile state including a threshold value to trigger a filter parameter associated with the profile objective, where the filter parameter comprises a content type to be processed so as to facilitate achieving the profile objective. The filter parameter is determined based on the profile state surpassing the user profile criterion. A content recognition model is used to identify a message content of each message. An electronic message that matches the content type is identified and filtered at the electronic messaging client to prevent the content from being presented.

Abstract: This power supply and demand planning device includes: an output range calculation unit that calculates the output range of a power generator that satisfies a plurality of restriction conditions; a power generation output range output calculation unit that calculates the power generator output in a single cross section on the basis of the calculated output range calculated; and a past specified cross section output correction unit that calculates a target output in the single cross section when a restriction condition violation occurs in the power generator output in the calculated single cross section calculated, and in order to eliminate a restriction condition violation, corrects the output range and the power generator output in the single cross section and a past cross section further in the past than the single cross section so that the power generator output in the single cross section becomes the target output.

Abstract: An autonomous storage and retrieval tower for receiving, storing, and dispensing free physical products.

Abstract: A system and method for analyzing coverage, bias and model explanations in large dimensional modeling data includes discretizing three or more variables of a dataset to generate a discretized phase space represented as a grid of a plurality of cells, the dataset comprising a plurality of records, each record of the plurality of records having a value and a unique identifier (ID). A grid transformation is applied to each record in the dataset to assign each record to a cell of the plurality of cells of the grid according to the grid transformation. A grid index is generated to reference each cell using a discretized feature vector. A grid storage for storing the records assigned to each cell of the grid is then created. The grid storage using the ID of each record as a reference to each record and the discretized feature vector as a key to each cell.

Abstract: Implementations of this disclosure provide an anomaly detection system and methods of performing anomaly detection on a time-series dataset. The anomaly detection may include utilization of a forecasting machine learning algorithm to obtain a prediction of points of the dataset and comparing the predicted value of a point in the dataset with the actual value to determine an error value associated with that point. Additionally, the anomaly detection may include determination of a sensitivity threshold that impacts whether points within the dataset associated with certain error values are flagged as anomalies. The forecasting machine learning algorithm may implement a seasonality component determination process that accounts for seasonality or patterns in the dataset. A search query statement may be automatically generated through importing the sensitivity threshold into a predetermined search query statement that implements that forecasting machine learning algorithm.

Abstract: The present invention relates to systems and methods for intelligently optimizing supply chain is provided. In particular, the systems and methods provide the capability to configure supply chain systems so as to: balance between cost and service is optimised and profitability maximised; configure system parameters to respond to both current and future risks; ensure that variability is built into plans enabling maximised efficiency; and human error and bias are eliminated from the planning process such that pro-active rather than reactive behaviour becomes the norm.

Abstract: A system, method and computer readable medium for dynamically determining location-based rules affecting item sales at POS registers for a particular location and for distributing the location-based rules to the POS registers is discussed. A location-based ruleset of rules regulating sales of items at an individual store is created and transmitted to the store along with a rule group to item map file that maps rules and warnings to individual store items. A rules engine at the store provides updated item sale rules and warnings to POS registers based on the location-based ruleset and rule group to item map file. Embodiments thus provide an improved system for updating POS registers at different store locations with location-based rules specific to the store location.

Abstract: Embodiments of the invention are directed to methods and devices for predicting interactions. One embodiment is directed to a method comprising receiving, by one or more computers, interaction data for a plurality of known interactions between resource providers and users, and creating a topological graph based on the plurality of known interactions. The method may further comprise determining, by the one or more computers, a plurality of communities to form a predictive model, and receiving a request for a prediction. In addition, the method may comprise applying the request to the predictive model, by the one or more computers, by identifying a community in the plurality of communities corresponding to the request, determining a node within the identified community, and providing information regarding the node as the requested prediction.

Abstract: A method is provided for performing dynamic inferencing at a node configured to communicate with other nodes of an IoT hierarchy. In the method, a schema for an asset object associated with one or more of at least one physical process and at least one physical device is received at the node. The schema is formatted according to an inferencing engine model format. An artificial intelligence model capable of being executed in an inferencing engine is received at the node. Data indicative of one or more of a current state of at least one physical process and a current state of at least one physical device is received at the node. The received data according to the schema and an inferencing engine are processed at the node. The inferencing engine generates a new predictive attribute based on the set of attributes, and the processing normalizes the received data according to the schema to generate normalized data, the normalized data includes the predictive attribute from the inferencing engine.

Abstract: A method for generating forecasts associated with items includes: obtaining a warm start forecasting request from a user; obtaining an item associated with the warm start forecasting request; obtaining item data associated with the item; making a determination that the item is a warm start item; in response to the determination: performing similarity calculations between the warm start item and additional items included in an item repository; identifying a portion of the additional items that are similar to the warm start item based on the similarity calculations; aggregating forecasts associated with the portion of the additional items to generate a warm start forecast; and providing the warm start forecast to the user.

Abstract: In order to facilitate machine learning for prediction using distinct dataset types, systems and methods include collecting content information from archived websites databases. Collecting historical event information from online sources, where the historical event information is associated with a plurality of historical events. Generating event-dependent products training datasets based on the content information and the historical event information, where the event-dependent content training datasets defines for content historical events that are associated with attributes of the content, attribute change of the content, or both. Training an attribute prediction machine learning model based on the event-dependent content training datasets. Applying the trained attribute prediction machine learning model to additional event information to predict, for content, a future attribute estimate, a future attribute change estimate, or both.

Abstract: Methods, systems, and computer storage media for providing a data analytics index associated with a crisis-recovery data analytics engine in a data analytics system. The data analytics index is a consolidated single index representation of a set of variables associated with a set of consumer behaviors. The crisis-recovery data analytics engine supports generating the data analytics index associated with a pre-crisis period and a crisis-recovery period. In operation, a crisis-recovery dataset—associated with a set of variables of a set of consumer behaviors that support quantifying recovery from a crisis event—is accessed. The set of consumer behaviors are selected based on a crisis-recovery machine learning model that is trained on a pre-crisis dataset and a crisis dataset for selecting the set of consumer behaviors. A data analytics index is generated based on the crisis-recovery dataset. A data visualization comprising crisis-recovery data indicating recovery from the crisis event is generated.

Abstract: Examples of the present disclosure describe systems and methods for probabilistic neural network architecture generation. In an example, an underlying distribution over neural network architectures based on various parameters is sampled using probabilistic modeling. Training data is evaluated in order to iteratively update the underlying distribution, thereby generating a probability distribution over the neural network architectures. The distribution is iteratively trained until the parameters associated with the neural network architecture converge. Once it is determined that the parameters have converged, the resulting probability distribution may be used to generate a resulting neural network architecture. As a result, intermediate architectures need not be fully trained, which dramatically reduces memory usage and/or processing time.

Abstract: As described herein, a system, method, and computer program are provided for handling business agreement updates requiring orchestration when an orchestration system is unavailable. A signed update to a business agreement requiring orchestration by an orchestration system is received. It is determined that the orchestration system is unavailable to orchestrate requirements of the signed update to the business agreement. One or more actions are performed to handle the signed update in response to determining that the orchestration system is unavailable.

Abstract: A system for managing consumer packaged goods (CPGs) is disclosed. The system includes a front end and a back end. The front end includes a mobile, airborne platform equipped with a digital image capturing device, and a wireless communications device. The backend is in communication with the front end via the wireless communications device, and includes a combinational convolutional neural network which derives models based on input data, a convolutional neural network which generates perception scoring utilizing input put from said combinational convolutional neural network, and a recurrent neural network which makes behavior predictions based on input from said convolutional neural network. The front end captures images of CPGs on a shelf, wherein the CPGs are subject to depletion over time, and wherein the backend generates predictions regarding the depletion state of the CPGs on the shelf as a function of time.

Abstract: Certain aspects of the present disclosure provide techniques for generating a metric, include receiving a rule defining one or more text strings; determining a set of transactions based on a user attribute; determining a first subset of transactions; determining a second subset of transactions; generating a first categorical distribution based on each transaction of the first subset of transactions being associated with a transaction description containing at least one text string of the one or more text strings; calculating a first unity metric based on the first categorical distribution; generating a second categorical distribution based on each transaction of the second subset of transactions being associated with a transaction description that does not contain a text string of the one or more text strings; calculating a second unity metric based on the second categorical distribution; determining a reliability metric for the rule; and providing the reliability metric.

Abstract: In an embodiment, an automated computer-based method for improving a computer system to be able to predict rejections of requisitions submitted to the computer system, the method comprising receiving a requisition from a client device; determining, at a scoring unit of a computer system, a probability value indicating a likelihood that the requisition would be rejected if the requisition is submitted to a requisition approval chain; transmitting the probability value from the computer system to the client device to be displayed on a display of the client device.

Abstract: Embodiments are directed to a machine learning engine that determines training documents and validation documents from a plurality of documents. The machine learning engine may determine attributes associated with the documents. In response to receiving a request to predict attribute values of a selected document the machine learning engine may train a plurality of ML models to predict the attribute values based on the training documents and the attributes and associate the trained ML models with an accuracy score. The machine learning engine may determine candidate ML models from the trained ML models based on the training accuracy scores. The machine learning engine may evaluate and rank the candidate ML models based on the request and the validation documents. The machine learning engine may generate confirmed ML models based on the ranked candidate ML models such that the confirmed ML models may answer the request.

Abstract: In certain aspects, the disclosed implementations include methods and systems for dynamically generating and providing transaction affinity recommendation data. In certain implementations, the transaction affinity recommendation data may include information that identifies a target merchant and associated merchant promotion data that may be generated based on a dynamic analysis of transaction data corresponding to an account record. The disclosed implementations may determine temporal relationships between pairs of sequential transactions corresponding to the account record, and based on such relationships, may identify pairs of transactions involving separate merchants. The disclosed implementations may determine affinity relationships associated with one or more pairs of the temporally related transactions and a target merchant, and based on such relationships, generate and provide the transaction affinity recommendation data.

Abstract: The present disclosure involves, systems, software, and computer-implemented methods for determining a location for a curbside branch. One example system comprises a memory storing instructions; at least one hardware processor interoperably coupled with the memory, wherein the instructions instruct the at least one hardware processor to: obtain historical and current exchange data of users associated with a particular location; obtain external data about the particular location; enrich the historical and current exchange data with the external data; identify one or more analysis rules to apply to the enriched data to determine an estimated demand for services associated with deployment of a mobile site at the particular location; compare estimated demand to a predetermined threshold; and in response to a determination that the estimated demand exceeds the predetermined threshold, identify the particular location for mobile site deployment.

Abstract: Systems and methods of the present disclosure enable the prediction of activity distribution across mediums of activities by employing processors that receive an activity record history across channels of mediums of activity. A candidate activity channel profile of a future activity channel is received that includes candidate channel attributes including: a medium attribute identifying the medium, and an activity category attribute identifying a category of activities of the future activity channel. Cannibalization forecasting models are used to predict channel metrics based on the candidate channel attributes and model parameters trained on the activity record history. New candidate channel attributes are automatically generated that improve the at least one channel metric based on the at least one channel metric, and a new candidate activity channel profile is automatically modified using the new candidate channel attributes. The channel metrics are displayed on computing devices associated with users.

Abstract: A method and system for time series forecasting are described. Forecasting a time series, though uncertain, has a potential to improve decision-making. It gives a picture of what can probably be expected. A live time series typically has trend and seasonality as its innate characteristic features, randomness being obvious in real-time. Described embodiments address complexity in configuring zero-configuration workflow forecasting methods at production scale. Described embodiments address complexity in configurating domain independent forecasting methods at production scale.

Abstract: Methods and systems for forecasting demand for a plurality of items are provided. In particular, the demand forecasting system and methods described herein are useful for predicting demand of products in a retail context. Forecast models are built and used to score incoming sales data to predict future demand for items. Forecast models are validated by evaluating actual demand against predicted demand and using that information to inform how future ensemble forecast will be generated. Forecasts may be broken down into smaller components to satisfy a variety of requests for data from client applications.

Abstract: Systems and methods for forecasting commercial real estate lease rent rates are disclosed. Exemplary embodiments receive a series of independent variables that represent attributes of a specific commercial real estate property, scale and normalize the series of independent variables, and assemble the scaled and normalized series of independent variables. Weightings may be applied to the assembled scaled and normalized series of independent variables to predict a rental rate for the specific commercial real estate property.

Abstract: A data processing method and apparatus, an electronic device, and a storage medium, relating to the field of computer technology. The method comprises: determining a coordinate order of each user according to video data (S110); obtaining a track of the user within a preset time interval according to the coordinate order (S120); determining, by means of the track, a relationship point of two shelves where the user stays (S130); and calculating, by means of the relationship point, a directional relationship diagram between shelves, so as to determine distribution positions of commodities in the shelves by means of the directional relationship diagram (S140).

Abstract: A supervisory computer is used with an energy management system of an electric vehicle. The energy management system includes a battery system having a plurality of battery subsystems. One of the battery subsystems is an abnormal battery subsystem while the remaining battery subsystems are normal battery subsystems. The supervisory computer includes at least one processor and at least one non-transitory computer-readable medium. The processor monitors the remaining state of charge, capacity, and resistance of the battery system and monitor the remaining state of charge and capacity of the abnormal battery subsystem, calculate a low integration bound value, calculate a remaining energy value for all of the normal battery subsystems with respect to the low integration value, calculate a remaining energy value for the abnormal battery subsystem, and summate the remaining energy values of the normal and abnormal battery subsystems to determine a global remaining energy value for the battery system.

Abstract: The present invention relates to systems and methods for model scoring and selection. Six or more metrics that are relevant to the model are initially selected, and weights are assigned to each metric. A first subset of the metrics are selected, including metrics for model fit and model error for primary regression. A second subset of metrics including at least two penalty functions are then selected for percentage of incidence. The scores from the primary regression and penalty calculations are aggregated into a final score. Multiple models can be scored and utilized to select a “best” model via an iterative culling of low scoring models and “breeding” of the high scoring models.

Abstract: Techniques are described herein for fleet electrification management. A method includes determining a composition of electric vehicles (EVs) to replace at least a portion of non-electric vehicles in a vehicle fleet while satisfying travel requirements of the vehicle fleet. The method includes estimating an energy demand of the composition of EVs. The method includes determining an electric vehicle supply equipment (EVSE) charging infrastructure to meet the estimated energy demand. The method includes providing one or more recommendations including at least one of: a fleet electrification recommendation for transitioning the vehicle fleet into the composition of EVs, or a charging infrastructure recommendation for implementing the EVSE charging infrastructure.

Abstract: Information can be provided in a convenient manner to help a user decide what events to attend. The information can include information regarding a popularity of the events. The popularity of the events can be based upon the number of people expected to attend, the number of the user's friends expected to attend, and/or the number of people fitting user-defined criteria who are expected to attend. The events for a user-specified time period can be shown on a map along with information representing the popularity of the events. Thus, the user can choose to attend those events in the user's area which are more popular. The map can be displayed on a user device, such as a cellular telephone. In this manner, the user can more readily attend those particular events which the user is more likely to enjoy.

Abstract: A system and method may retrieve transaction data for a plurality of electronic payment system transactions. The transaction data may describe a plurality of electronic payment transactions between a payment network system and a plurality of merchants. The transaction data may include a transaction location and a merchant store number. The system and method may parse the transaction data to identify the transaction location and the merchant store number and analyze the retrieved and parsed transaction data to identify one or more centroids for at least a portion of the retrieved and parsed transaction data. The one or more centroids may correspond to a merchant location for each transaction of the retrieved and parsed transaction data. The system and method may then clean the retrieved and parsed transaction data having a transaction location outside a threshold distance from the one or more centroids to include accurate merchant store number data.

Abstract: A method for estimating movements of a crowd between a first and second subregion in an area are monitored by a wireless sensor network. The wireless sensor network includes nodes configured to exchange a radio frequency signal through a first respective second link. The first respective second link crosses the first respective second subregion. The method includes the steps of exchanging radio frequency signals over the first and second link; and measuring respective first and second attenuations of the exchanged radio frequency signals over the first respective second link; and estimating based on a change in the attenuations a flow of the crowd between the first and second subregion. The estimating further includes estimating based on the first and second attenuations a density of the crowd in the first respective second subregion; and estimating based thereon a flux of the crowd between the first and second subregion.

Abstract: An integrated energy optimizer having an edge side and a cloud side. The edge side may incorporate an energy optimizer, a building management system connected to the energy optimizer, a controller connected to the building management system, and equipment connected to the controller. The cloud side may have a cloud connected to the energy optimizer and to the building management system, and a user interface connected to the cloud. Data from the field sensor may go to the optimizer and the building management system. The data may be processed at the optimizer and the building management system for proper settings at the building management system.

Abstract: Provided is a method for computer-implemented determination of a data-driven prediction model. The method processes digital input data having input and output variables and being semantically annotated based on a digital semantic representation having a hierarchical tree structure where each tree in the structure represents an input variable of the data, the leaf nodes of the respective tree being the discrete values of the input variable. The method of the embodiment provides a recoding of those discrete values by cutting off hierarchical levels of the respective trees. Based on this recoding, a plurality of data modifications is determined for the input data. Those data modifications are trained by a machine learning method where the trained machine learning method with the highest prediction quality is derived from the trained machine learning methods.

Abstract: A system and method for autogenerated synthetic operational customer satisfaction scoring and analysis, that generates synthetic calls and follows customer interactions across multiple channels as a single journey, which monitors synthetic calls for a plurality of failure events that may represent infrastructure related operational deficiencies that may adversely affect customer satisfaction and sentiment, which assigns a synthetic operational customer satisfaction score for each synthetic call, and which uses machine learning to optimize infrastructure configurations to improve customer satisfaction and sentiment.

Abstract: In accordance with some embodiments, a method for generating sentimental information associated with media content is provided, the method comprising: receiving user input data corresponding to a user; identifying a portion of the media content item based on the user input data; determining a sentiment based on the user input data, wherein the sentiment is one of a positive sentiment which indicates that the media content item was liked by the user or a negative sentiment which indicates that the media content item was disliked by the user; determining an amount of the sentiment based on the user input data; associating the amount of the sentiment with the portion of the media content item; and generating, for the media content item, sentimental information that indicates the amount of the sentiment associated with at least one portion of the media content item.

Abstract: A system is provided that is capable of processing clinical trial data from one or more sources, both structured and unstructured data, and integrate such data in a database for access by users. Further, such trial data is combined with other types of data to train an AI-based model, for the purpose of determining a probability of a particular trial outcome, among other insights. Further, the system can provide insights into optimizing this probability, for example by optimizing elements of the clinical trial design. Further, the system stores the acquired data from multiple sources and provides search, comparison, and reporting capability within a single interface.

Abstract: A method of training a neural network to approximate a forecasting error of a passenger-demand forecasting model that includes calculating, using the forecasting model, a historical passenger demand forecast for each key level in a set of key levels and for each departure date in a set of historical departures dates; applying a dropout model to the historical passenger demand forecasts to create a training sample; training, using the historical passenger demand forecasts and the training sample, the neural network, to approximate forecasting errors associated with the forecasting model; calculating, using the forecasting model, a future passenger demand forecast for each key level in the set of key levels and for each departure date in a set of future dates; and approximating, using the trained neural network, the forecasting error associated with the future passenger demand forecasts for the second set of departure dates.

Abstract: Disclosed herein are a system and method for recommending repeat-purchase products using modified self-similarity. The system is operated by a computer in an environment in which user equipment (UE), a database, and a web server are connected over a network.