Abstract: Embodiments of the present disclosure are directed to specific, networked, and technological processes and operations. For example, various embodiments are directed to a distributed and collaborative planning tool that allows users to create events based on a presently-approved production plan, enable/disable events submitted by other users, submit proposals, approve proposals, and integrate proposals to update the presently-approved production plan. An event published by a first user is communicated to other users using a single, universal framework that tracks and communicates proposed changes between remote users and their associated client devices during the planning cycle without the need to store multiple iterations of full-size production plans.

Abstract: An out-of-stock indicator is received that indicates a product is out-of-stock or believed to be out-of-stock. Information about the features of the products and store are obtained or determined. The features are applied to a first machine learning model, which yields a probability that the item is out-of-stock. The obtained probability is compared to a threshold, and if the probability value is above a threshold, then the PI value is adjusted. If not above the threshold, then scans are monitored for out-of-stock conditions, and some time later the features will be applied to a different model, and the above-process repeated. In aspects, this process occurs over a certain time period or until the PI is adjusted.

Abstract: Systems, apparatuses, and methods are provided herein for in-store production management. A method comprises training a probabilistic demand model using the training dataset based on a machine learning algorithm, determining a probabilistic forecast of a demand for a perishable product at a store location on a date based on the probabilistic demand model, applying an objective function to each of the plurality of demand values in the probabilistic forecast to determine an objective value for each of the plurality of demand values, determining a target supply value based on a select demand value with a highest objective value, determining a production plan based on the target supply value and a carryover value retrieved from the store data database, and providing a production management user interface that comprises a display of an identifier of the perishable product and the production plan associated the perishable product.

Abstract: Machine learning segmentation methods and systems that perform segmentation quickly, efficiently, cheaply, and optionally provides an interactive feature that allows a user to alter the segmentation until a desired result is obtained. The automated machine learning segmentation tool receives all potentially important attributes and provides segmentation of items. It also receives information about important features of the data and finds how best to differentiate between groups using cluster-based machine learning algorithms. In addition, visualization of the segmentation explains to a user how the segmentation was obtained.

Abstract: In a cloud computing environment, a configurable transaction status interface of an enterprise computing platform enables application developers to customize applications efficiently. The configurable transaction status interface includes a global variable that is populated using a method for implementing a configurable status map that maps a transaction result to the global variable without having to customize the application. The configurable transaction status interface allows third party vendors of transaction gateways to use the configurable status map to map one or more of their gateway-specific transaction results to a generic status defined in the global variable. The vendors manage the configurable status map in their own namespace independently of the application with which they are interfacing.

Abstract: A design management apparatus (100) includes a conversion unit (12) and a test unit (13). The conversion unit (12) generates model information (403) that is a format of design information (303) being converted, the design information (303) being the design information (303) created in a process of mechanical design in an engineering chain, and generates model information (405) that is a format of design information (305) being converted, the design information (305) being the design information (305) created in a process of control design in the engineering chain. The test unit (13) associates the model information (403) and the model information (405) using entire reference information (22) that associates the model information (403) and the model information (405), and tests for consistency between the model information (403) and the model information (405) associated.

Abstract: Various embodiments provide systems, methods, and computer program products for dynamically identifying and engaging at least one on-demand packaging customer in a consolidated fashion. The system comprises one or more computer processors configured to: receive qualification data associated with at least one customer; calculate whether the at least one customer is a qualified customer; in response to identifying the qualified customer, determine whether the qualified customer is a targeted customer; in response to identifying the targeted customer, evaluate a readiness of at least one facility operated by the targeted customer for engagement of an on-demand packaging configuration; in response to determining a positive facility readiness, generate one or more documents for execution by at least the targeted customer, the carrier, and the supplier; and upon execution of the one or more documents, facilitate implementation of the on-demand packaging configuration at the at least one facility.

Abstract: A system and method may be provided for associating bio-signal data (e.g. EEG brain scan data) from at least one user with at least one music data item (e.g. song, or piece of music). By associating bio-signal data, or emotions determined therefrom, with music, the system may establish a data store of music associated with emotions. That database may then be leveraged upon determining that a user is feeling a particular emotion through an EEG scan. When a particular emotion is detected in EEG data of a user, the system may then respond based at least partly on the same or similar emotion being associated with one or more music data items in the system. For example, the system may recommend a particular song associated with the same emotion presently being experienced by the user.

Abstract: In one aspect, a computerized method of computer vision based dynamic universal fashion ontology fashion rating and recommendations includes the step of receiving one or more user-uploaded digital images. The method includes the step of implementing an image classifier on the one or more user-uploaded digital images, to classify a set of user-uploaded fashion content of the one or more user-upload digital images. The method includes the step of receiving a set of fashion rules input by a domain expert. The set of rules determine a set of apparel to match with the set of user-uploaded fashion content, generating a dynamic universal fashion ontology with the image classier and a text classier. The dynamic universal fashion ontology comprises an ontology of set of mutually exclusive attribute classes. The method includes the step of using the dynamic universal fashion ontology to train a specified machine learning based fashion classifications.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating a recommended target audience based on determining a predicted attendance utilizing a neural network approach. For example, the disclosed systems can utilize an approximate nearest neighbor algorithm to identify individuals that are within a similarity threshold of invitees for an event. In addition, the disclosed systems can implement an attendance prediction model to determine a probability of an invitee attending the event. The disclosed systems can further determine a predicted attendance for an event based on the individual probabilities. Based on identifying the similar individuals to, and the attendance probabilities for, the invitees, the disclosed systems can generate a recommended target audience to satisfy a target attendance for an event based on a predicted attendance for the event.

Abstract: Systems and methods for features engineering, in which internal and external signals are received and fused. The fusing is based on meta-data of each of the one or more internal signals and each of the one or more external signals. A set of features is generated based on one or more valid combinations that match a transformation input, the transformation forming part of library of transformations. Finally, a set of one or more features is selected from the plurality of features, based on a predictive strength of each feature. The set of selected features can be used to train and select a machine learning model.

Abstract: A method for predictively updating one or more user parameters associated with a user of a learning system includes predicting, based on the one or more user parameters, a predicted activity of the user, receiving an actual activity of the user, comparing the predicted activity to the actual activity, and updating the one or more user parameters in response to determining that the predicted activity does not match the actual activity. The method may further include scheduling one or more learning interactions based on the one or more updated learning parameters, where the scheduling includes selecting at least one of a timing of the one or more learning interactions or a type of the one or more learning interactions.

Abstract: A method comprising, by a processing unit and a memory: obtaining a training set of data; dividing sets of data into a plurality of groups, wherein all sets of data for which feature values meet at least one similarity criterion, are in the same group, storing in a reduced training set of data, for each group, at least one aggregated set of data, wherein, for a plurality of the groups, a number of aggregated sets of data is less than a number of the sets of data of the group, wherein the reduced training set of data is suitable to be used in a classification algorithm for determining a relationship between the at least one label and the features of the electronic items, thereby reducing computation complexity when processing the reduced training set of data, compared to processing the training set of data.

Abstract: A processor may receive workflow data associated with an environment having one or more smart devices. A processor may analyze the workflow data to identify one or more activities associated with the environment. A processor may generate an intelligent environment for one or more users using one or more intelligent environment (IE) devices. The one or more IE devices may be configured to collect user feedback from one or more users. A processor may modify at least one of the one or more activities in the environment based, at least in part, on the user feedback.

Abstract: Aspects of the present invention provide devices that identify a question in a text message of a chat session between a plurality of computing devices, wherein each of the plurality of computing devices includes an identifier of a user of a corresponding computing device in the text message, analyze digital data to determine a relationship between the users of the plurality of computing devices, construct search parameters for digital content external to the chat session according to the identified question and determined relationship, search for the digital content over a network according to the constructed search parameters to obtain search results, rank the search results according to the determined relationship, and distribute the ranked search results according to the determined relationship to at least one of the plurality of computing devices in an overlay of the chat session.

Abstract: Described are systems and processes for generating dynamic estimated time of arrival predictive updates for delivery of perishable goods. In one aspect a system is configured for generating dynamic estimated time of arrival (ETA) predictive updates between a series of successive events for real-time delivery of orders. For each order, a plurality of delivery events and corresponding timestamps are received from devices operated by customers, restaurants, and couriers. Based on the timestamps, the system generates a plurality of ETA time predictions for one or more of the delivery events with trained predictive models that use weighted factors including historical restaurant data and historical courier performance. As additional timestamps are received for a delivery event, the trained predictive models dynamically update the ETA time predictions for successive events. The predictive models may be continuously trained by updating the weighted factors based on the received timestamps.

Abstract: A method of smart urban public transport management, an Internet of Things system, and a storage medium are provided. The method of smart urban public transport management is implemented by the public transport management platform. The method of smart urban public transport management includes: obtaining the predicted pedestrian flow at a plurality of places through the place management platform; determining the target places where the predicted pedestrian flow is greater than the pedestrian flow threshold based on the predicted pedestrian flow at the plurality of places; obtaining the bus lines passing by the target places and adjusting the departure frequency of the bus lines. The Internet of Things system includes a user platform, a service platform, a public transport management platform, a sensor network platform, and an object platform. The method can be executed after the computer instructions stored in the computer-readable storage medium are read.

Abstract: Provided are various mechanisms and processes for generating dynamic merchant similarity predictions. In one aspect, a system is configured for receiving historical datasets that include a series of merchants from historical browsing sessions generated by one or more users. The merchants are converted into corresponding vector representations for training a predictive model to output associated merchants based on a generated weighted vector space. Once sufficiently trained, data from a new browsing session may be received, which may include a target merchant. The target merchant is input into the predictive model as a vector to output one or more context merchants having vectors with the highest cosine similarity value to the target merchant vector. Selected context merchants may then be transmitted to the user device as targeted merchant suggestions in the new browsing session. The predictive models may be continuously trained using data received from subsequent browsing sessions.

Abstract: Product demand forecasting accuracy utilizes partitional clustering of time series data with dynamic time warping. The product demand forecasting disclosed herein is particularly suited to forecasting product demand for products with limited sales data. Time-series sales data of a producs (or group of products) with limited sales data (e.g. a sparse or no time series of sales data) are dynamically time warped with sales data of products, or groups of products, having extensive sales data (e.g., an extensive time series of sales data) to determine a clustering model with an optimal number of clusters and a prototype time series for each cluster in the model. The prototype time series for the cluster in which the product (or group of products) with limited sales data lies is utilized as its product demand forecast.

Abstract: The subject technology identifies a series of journey event types in an online user journey, the event types including an impression event, an email event, a click event, and a website visit, and assigns an encoder to each event type. Using an assigned encoder, the technology encodes each event type to generate an encoded vector for each event type. The encoded vector is representative of at least a portion of the online user journey relating to that event type. The technology generates an encoded vector for each event type to create a set of encoded vectors, the set of encoded vectors including one or more of an impression event encoded vector, an email event encoded vector, a click event encoded vector, and a website visit encoded vector.

Abstract: A system, method, and apparatus provide the ability to generate and deliver personalized digital content. Multiple content tests are performed by presenting different variants of content to a set of different consumers of one or more consumers. A machine learning (ML model is generated and trained based on an analysis of results of the multiple content tests. Based on the ML model, personalization rules, that specify a certain variance for a defined set of facts, are output. The personalization rules are exposed to an administrative user who selects one or more of the personalization rules. A request for content is received from a requesting consumer. Based on similarities between the defined set of facts and the requesting consumer, a subset of the selected personalization rules are selected. The content is personalized and delivered to the requesting consumer based on the further selected personalization rules.

Abstract: The present disclosure provides systems, methods, and metrics that filter out online visitor behavioral data that represents a potential lead with a high likelihood to convert to a vehicle sale from online visitor behavioral data that does not represent a potential lead with a low likelihood to convert to a vehicle sale, based on a mapping of sales back to observed website and vehicle configurator data. This enables more effective lead generation and the more efficient targeting of online incentive offers and sales “nudges,” for example. Further, the present disclosure enables web analytics data to be combined with sales data for sales forecasting in general.

Abstract: Embodiments of the present invention are directed to facilitating performing online data decomposition. In accordance with aspects of the present disclosure, an incoming data point of a time series data set is obtained. Thereafter, an iterative process of estimating trend and seasonality is performed to decompose the incoming data point to a set of data components based on a particular set of previous data points of the time series data set and corresponding data components. Generally, the set of data components for the incoming data point include a trend component, a seasonality component, and a residual component. The set of data components is provided for analysis of the incoming data point, such as, for example, to identify data anomalies.

Abstract: A system and method for automated order preparation and fulfillment timing. The system is a cloud-based network containing an optimization server, portals for restaurants, customers, and drivers to enter their information, and an optimization engine which determines an optimal state for consumption of, and timing for fulfillment of, an order based on a multitude of variables associated with the business enterprises and delivery driver availability. The system may be accessed through web browsers or purpose-built computer and mobile phone applications.

Abstract: Disclosed are systems, methods, and non-transitory computer-readable media for a geographic recommendation platform. The geographic recommendation platform receives data identifying a geographic region specified by a user and gathers data relating to the geographic region. The geographic recommendation platform determines, based on the data relating to the geographic region, an anticipated demand for an item within geographic region. The anticipated demand indicates how likely the item is to be purchased by a user that is located within the geographic region. The geographic recommendation platform generates a recommendation for the item based on the anticipated demand. The recommendation indicates the anticipated demand for the item within the geographic region. The geographic recommendation platform transmits the recommendation to the user.

Abstract: Methods, apparatuses, and computer program products are described herein that are configured to be embodied as a benchmarking service. In an example, an apparatus is configured to access input data, wherein the input data is representative of a current project; parse the input data to generate one or more input project units; extract one or more features from the one or more input project units, wherein the features are representative of at least one of project statistics, project bugs, project releases, project documentations, and organization data; receive a benchmarking model, wherein the benchmarking model was derived using a historical data set; and generate an output based on the benchmarking model and the one or more features, wherein the output is configured to provide an evaluation of the current project in the form at least one of a score and one or more recommendations.

Abstract: An Artificial Intelligence (AI)-based attribute prediction system generates predictions for attributes of highly customized equipment in response to received user requests. Processed historical data is initially used to generate feature combinations which are then employed along with a plurality of statistical and machine learning (ML) models in order to identify a best scoring model-feature combination in two selection cycles using multiple selection criteria. The predictions for an attribute are generated by the best scoring model and feature combination. Various insights regarding the features affecting the attribute can be additionally derived to provide recommendations to the user.

Abstract: Various embodiments described herein relate to techniques for forecasting with state transitions and confidence factors. In this regard, a system is configured to segment data associated with one or more assets to determine a set of classifications for one or more attributes related to the one or more assets. The system is also configured to generate a state machine associated with a Markov chain model based on the set of classifications for the data. Furthermore, the system is configured to perform a machine learning process associated with the state machine to determine one or more behavior changes associated with the one or more attributes related to the one or more assets. The system is also configured to predict, based on the one or more behavior changes associated with the one or more attributes related to the one or more assets, a change in demand data for the one or more assets during a future interval of time.

Abstract: A method and a system for predicting and using customer lifetime value (CLV). The method include: providing a classifier trained using customer feature data during a first period of time as input and whether there is spending during a second period of time as classifier label; providing a regressor trained using the customer feature data during the first period of time as input and amount of spending during a second period of time as regressor label; performing the classifier using customer feature data during a third period of time to obtain customers having positive predicted classifier labels; and performing the regressor using the customer feature data during the third period of time for the customers having positive predicted classifier labels, to obtain CLVs of the customers.

Abstract: A method for predicting user purchase by a user of a first site includes: selecting a distribution representing a probability distribution (PD) of inter-purchase-times (IPTs) across the first site and a second other site for each user, assigning each purchase of each user to one of the first site and the second site according to a Stochastic model, combining the selected PD with the Stochastic model to generate a PD of IPTs for only the first online site, estimating parameters of the probability distribution of IPTs for the first site by applying a Statistical modeling approach to features of each user, applying a sequence of observed IPTs of a given user for the first site and the parameters of the given user to the selected distribution to generate a probability, and determining whether the next purchase occurs on the second site based on the probability.

Abstract: Techniques are described herein for providing adaptable testing and benchmarking of user experiences with respect to one or more products. In some embodiments, the techniques include systems and methods for predicting performance of facets of a user experience under new testing and benchmarking methodologies. The systems and methods may generate a prediction for the results of a UX test even though the methodology and mechanics to quantify the user experience may vary significantly from previous methodologies. The techniques allow for methodologies to evolve over time without losing historical context or the ability to meaningfully compare historical test results with tests run using updated testing and benchmark models. Further, the techniques allow for benchmarks to be computed in real-time or near real-time as methodologies change without requiring tests to be run according to the new methodologies.

Abstract: Digital content interaction prediction and training techniques that address imbalanced classes are described. In one or more implementations, a digital medium environment is described to predict user interaction with digital content that addresses an imbalance of numbers included in first and second classes in training data used to train a model using machine learning. The training data is received that describes the first class and the second class. A model is trained using machine learning. The training includes sampling the training data to include at least one subset of the training data from the first class and at least one subset of the training data from the second class. Iterative selections are made of a batch from the sampled training data. The iteratively selected batches are iteratively processed by a classifier implemented using machine learning to train the model.

Abstract: Techniques are described for generating seasonal forecasts. According to an embodiment, a set of time-series data is associated with one or more classes, which may include a first class that represent a dense pattern that repeats over multiple instances of a season in the set of time-series data and a second class that represent another pattern that repeats over multiple instances of the season in the set of time-series data. A particular class of data is associated with at least two sub-classes of data, where a first sub-class represents high data points from the first class, and a second sub-class represents another set of data points from the first class. A trend rate is determined for a particular sub-class. Based at least in part on the trend rate, a forecast is generated.

Abstract: Embodiments provide techniques for autonomous vehicle fleet modeling and simulation, such as within a dynamic transportation matching system utilizing one or more vehicle types such as non-autonomous vehicles and autonomous vehicles. An autonomous fleet simulation model may be generated based on real-world parameters of an autonomous vehicle fleet, and the parameters may be modified in a simulation in order to determine optimized values that may be applied to the real-world autonomous vehicle fleet.

Abstract: This disclosure relates to a system and method to estimate demand transfer of a product while considering performance of all the products of a category simultaneously. It would be appreciated that the demand of a removed product transfers to other products of same category in a store. In addition the demand transfer is influenced by sales drivers such as product level promotion and competitor prices, store location, weather and seasonality. By considering these factors the proposed approach provides a method to estimate demand transfer of a product. It is addressed by creating multivariate multi structure machine learning models and estimating demand transfer values by using suitable scenario generator for product availability. It enables to estimate more holistic demand transfer values by simultaneous consideration of individual product behaviours with respect to other products availability and other sales drivers.

Abstract: Techniques for using online engagement footprints for video engagement prediction are provided. In one technique, events are received from multiple client devices, each event indicating a type of engagement of a video item from among multiple types of engagement. One or more machine learning techniques are used to train a prediction model that is based on the events and multiple features that includes the multiple types of engagement. In response to receiving a content request, multiple entity feature values are identified for a particular entity that is associated with the content request. Two or more of the entity feature values correspond to two or more of the types of engagement. A prediction is generated based on the entity feature values and the prediction model. The prediction is used to determine whether to select, from candidate content items, a particular content item that includes particular video.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to project ratings for future broadcasts of media. Disclosed example methods include normalizing, with a processor, audience measurement data corresponding to media exposure data, social media exposure data and programming information associated with a future quarter to determine normalized audience measurement data. Disclosed example methods also include classifying a media asset based on the programming information to determine a media asset classification.

Abstract: Item delivery optimization may be provided. Items may be identified as available for delivery at a first location to at least one second location. A route from the first location to the second location may be calculated and at least a portion of the route from the first location to the second location may be provided to a delivery mechanism, such as an automated delivery device and/or a user.

Abstract: An example method of complementary consumer item selection includes: receiving, by a computer system, an image representing a first consumer item of a reference set of consumer items; identifying a target category of a complementary consumer item to be associated with the reference set of consumer items; generating, by a neural network processing the set of images, a feature embeddings representing the first consumer item in relation to the target category; selecting, using the feature embedding, from a set of available consumer items, a plurality of candidate consumer items associated with the target category; and selecting, among the plurality of candidate consumer items, the complementary consumer item to be associated with the reference set of consumer items.

Abstract: System and method to build and score predictive model for numerical attributes are provided. The system includes a memory and a processing subsystem. The processing subsystem is configured to select one or more numerical variables from the plurality of data sets based on a plurality of parameters, to apply feature engineering and transformation on the one or more numerical variables, to perform time series forecasting on the one or more numerical variables based on the plurality of features extracted, to evaluate and select appropriate prediction technique based a regression technique based on a plurality of elements, to build a prediction model, to score the built prediction model based on the performed time series forecasting and an evaluated regression technique and to predict the built prediction model based on an obtained score. Further, the system uses the plurality of parameters and the prediction method to score and predict the prediction model.

Abstract: This disclosure describes one or more implementations of a model segmentation system that generates accurate audience segments for client devices/individuals utilizing multi-class decision tree machine-learning models. For example, in various implementations, the model segmentation system generates a customized loss penalty matrix from multiple loss penalty matrices. In particular, the model segmentation system can generate regression mappings of model evaluation metrics for a plurality of decision tree models and combine loss penalty matrices based on the regression mappings to generate a customized loss penalty matrix that best fits an administrator's customized needs of segment accuracy and reach. The model segmentation system then utilizes the customized loss penalty matrix to train a multi-class decision tree machine-learning model to classify client devices into non-overlapping audience segments.

Abstract: Systems and methods are provided herein for producing and vending ice in an energy-efficient manner. A system for producing and vending ice comprises: an ice-making device configured to manufacture ice; a reservoir coupled to the ice-making device and configured to store the manufactured ice manufactured; and a control circuit communicatively coupled to the ice-making device, the control circuit configured to: determine a beginning ice inventory at a first time; determine a predicted ice demand for a first period, wherein the predicted ice demand is a function of at least a historical factor, a weather factor, a customer social event factor, and a public social event factor; determine an ice manufacturing quantity based on the beginning ice inventory, the predicted ice demand, and a manufacturing capacity of the ice-making device; and cause the ice-making device to produce an amount of ice consistent with the determined ice manufacturing quantity.

Abstract: An automatic system facilitates selection of media properties on which to display an advertisement, responsive to a profile collected on a first media property, where a behavioral-targeting company calculates expected profit for an ad correlated with the profile and arranges for the visitor to be tagged with a tag readable by the selected media property. The profit can be calculated by deducting, from the revenues that are expected to be generated from an ad delivered based on the collected profile, at least the price of ad space at a media property where the BT company might like to deliver ads to the profiled visitor. When the calculated profit is positive (i.e., not a loss), the BT company arranges for the visitor to be tagged with a tag readable by the selected media property through which the BT company expects to profit.

Abstract: A processing device may analyze a set of time series data using a time series forecasting model comprising an attributes model and a trend detection model. The attributes model may comprise a modified gradient boosting decision tree (GBDT) based algorithm. Analyzing the set of time series data comprises determining a set of features of the set of time series data, the set of features including periodic components as well as arbitrary components. A trend of the set of time series data may be determined using the trend detection model and the set of features and the trend may be combined to generate a time series forecast.

Abstract: Embodiments generate a model of demand of a product that includes an optimized feature set. Embodiments receive sales history for the product and receive a set of relevant features for the product and designate a subset of the relevant features as mandatory features. From the sales history, embodiments form a training dataset and a validation dataset and randomly select from the set of relevant features one or more optional features. Embodiments include the selected optional features with the mandatory features to create a feature test set. Embodiments train an algorithm using the training dataset and the feature test set to generate a trained algorithm and calculate an early stopping metric using the trained algorithm and the validation dataset. When the early stopping metric is below a predefined threshold, the feature test set is the optimized feature set.

Abstract: A system and method for en-route business selection, routing, and order preparation timing. The system is a cloud-based network containing an optimization server, portals for restaurants, customers, and drivers to enter their information, and an optimization engine which determines optimal pickup and delivery times for delivery drivers based on a multitude of variables associated with the business enterprises and delivery driver availability. The system may be accessed through web browsers or purpose-built computer and mobile phone applications.

Abstract: Methods and apparatus to estimate audience sizes using deduplication based on vector of counts sketch data are disclosed. An example apparatus to determine an audience size for media based on vector of counts sketch data includes: a coefficient analyzer to determine coefficient values of a polynomial based on variances, a covariance, and cardinalities corresponding to a first vector of counts from a first database and a second vector of counts from a second database; an overlap analyzer to determine a real root of the polynomial, the real root corresponding to an estimate of an overlap between the first vector of counts and the second vector of counts; and a report generator to estimate the audience size based on the estimate of the overlap and the cardinalities of the first vector of counts and the second vector of counts.

Abstract: Requests for computing resources and other resources can be predicted and managed. For example, a system can determine a baseline prediction indicating a number of requests for an object over a future time-period. The system can then execute a first model to generate a first set of values based on seasonality in the baseline prediction, a second model to generate a second set of values based on short-term trends in the baseline prediction, and a third model to generate a third set of values based on the baseline prediction. The system can select a most accurate model from among the three models and generate an output prediction by applying the set of values output by the most accurate model to the baseline prediction. Based on the output prediction, the system can cause an adjustment to be made to a provisioning process for the object.

Abstract: A system and method are disclosed including a computer and a processor and memory. The computer receives historical sales data comprising aggregated sales data for one or more items from one or more store for at least one past time period. The computer further trains a cyclic boosting model to learn model parameters by iteratively calculating for each feature and each bin factors for at least one full feature cycle. The computer further predicts one or more demand quantities during a prediction period by applying a prediction model to historical supply chain data, wherein a training period is earlier than the prediction period, and each of the one or more demand quantities is associated with at least one item of the one or more items and at least one stocking location of the one or more stocking locations during the prediction period and rendering a demand prediction feature explanation visualization.

Abstract: Systems and methods for features engineering, in which internal and external signals are received and fused. The fusing is based on meta-data of each of the one or more internal signals and each of the one or more external signals. A set of features is generated based on one or more valid combinations that match a transformation input, the transformation forming part of library of transformations. Finally, a set of one or more features is selected from the plurality of features, based on a predictive strength of each feature. The set of selected features can be used to train and select a machine learning model.