Abstract: Systems and methods for performing structure and attribute based graph partitioning. An exemplary system can receive a plurality of transactions, with each transaction identifying a point of sale and an item exchanged by the transaction. The system can generate a structure graph of the plurality of transactions and receive a request to partition the structure graph into sub-categories. The system can then retrieve, from a database, attributes of each respective structure vertex in the structure vertices. Using those attributes and the structure graph, the system can generate a structure-attribute graph. The system can then generate a unified neighborhood matrix of the structure-attribute graph; and partition the structure graph based on the unified neighborhood matrix.

Abstract: A system including one or more processors and one or more non-transitory computer-readable media storing computing instructions configured to run on the one or more processors and perform: receiving historical delivery records over a predetermined time period from partners associated with items offered to subregions through an online platform; generating nodes for combinations each comprising a respective one of the partners, a respective one of the items offered by the partners, and a respective one of the subregions; generating, using a machine learning model, a respective classification for each respective node on whether to tag the each respective node as deliverable in a predetermined time window; and automatically tagging a portion of the nodes as deliverable in the predetermined time window in the online platform. Other embodiments are disclosed.

Abstract: An example method for predicting average inventory with newly added items can include: aggregating sales data of a plurality of items, the items comprising training items and new items; identifying, using a set of predefined rules, a data set of similar items on the training items for each of the new items, the set of predefined rules comprising a first stage similarity module, a second stage similarity module, and a second stage classification module; obtaining target metrics for each of the new items, the target metrics being turn predictions from the data set of the similar items; calculating mean errors of the turn predictions to identify a set of turn predictions with mean errors lower than a dynamic threshold; obtaining an ultimate turn prediction for each new item by averaging the set of turn predictions; and predicting an average inventory for each new item based on the ultimate turn prediction.

Abstract: Examples provide a system for generating document-based responses to user provided queries. The response generation component creates a set of generated utterances based on text associated with a set of sentences in at least one document. Each utterance in the set of generated utterances is assigned an intent. A filter component identifies at least one utterance from the set of generated utterances having a same intent as a user-provided utterance to form a set of filtered utterances. A selection component identifies one or more utterance(s) from the set of filtered utterances having a shortest distance from the user-provided utterance. If more than one utterance is identified, a weighted summarization response is output to the user based on a predefined answer to each utterance in the selected set of utterances.

Abstract: Examples provide a system for generating document-based responses to user provided queries. The response generation component creates a set of generated utterances based on text associated with a set of sentences in at least one document. Each utterance in the set of generated utterances is assigned an intent. A filter component identifies at least one utterance from the set of generated utterances having a same intent as a user-provided utterance to form a set of filtered utterances. A selection component identifies one or more utterance(s) from the set of filtered utterances having a shortest distance from the user-provided utterance. If more than one utterance is identified, a weighted summarization response is output to the user based on a predefined answer to each utterance in the selected set of utterances.

Abstract: An example method for predicting average inventory with newly added items can include: aggregating sales data of a plurality of items, the items comprising training items and new items; identifying, using a set of predefined rules, a data set of similar items on the training items for each of the new items, the set of predefined rules comprising a first stage similarity module, a second stage similarity module, and a second stage classification module; obtaining target metrics for each of the new items, the target metrics being turn predictions from the data set of the similar items; calculating mean errors of the turn predictions to identify a set of turn predictions with mean errors lower than a dynamic threshold; obtaining an ultimate turn prediction for each new item by averaging the set of turn predictions; and predicting an average inventory for each new item based on the ultimate turn prediction.

Abstract: Systems and methods for performing structure and attribute based graph partitioning. An exemplary system can receive a plurality of transactions, with each transaction identifying a point of sale and an item exchanged by the transaction. The system can generate a structure graph of the plurality of transactions and receive a request to partition the structure graph into sub-categories. The system can then retrieve, from a database, attributes of each respective structure vertex in the structure vertices. Using those attributes and the structure graph, the system can generate a structure-attribute graph. The system can then generate a unified neighborhood matrix of the structure-attribute graph; and partition the structure graph based on the unified neighborhood matrix.

Abstract: In some embodiments, methods and systems of detecting and monitoring shoplifting of products by individuals at a retail facility include sensors coupled to a product and located at a fixed threshold distance relative to each other. The sensors are configured to generate an alert signal indicating that the fixed threshold distance between the first and second sensors has been altered. An alert sensor positioned in the product stocking area of the retail facility is configured to receive the alert signal from one or both of the product sensors. A central computing device at the retail facility is configured to determine the alert sensor that received the alert signal, and to activate a video camera in order to acquire a photo or video of the portion of the product stocking area of the retail facility associated with the alert sensor that received the alert signal.

Abstract: In some embodiments, methods and systems of detecting and monitoring shoplifting of products by individuals at a retail facility include sensors coupled to a product and located at a fixed threshold distance relative to each other. The sensors are configured to generate an alert signal indicating that the fixed threshold distance between the first and second sensors has been altered. An alert sensor positioned in the product stocking area of the retail facility is configured to receive the alert signal from one or both of the product sensors. A central computing device at the retail facility is configured to determine the alert sensor that received the alert signal, and to activate a video camera in order to acquire a photo or video of the portion of the product stocking area of the retail facility associated with the alert sensor that received the alert signal.

Abstract: Examples provide demand transference modeling for item assortment management. A demand prediction component analyzes item attribute data using a demand transference model to calculate a magnitude of demand transfer between items in a set of substitute items associated with a proposed item assortment. The proposed item assortment includes at least one assortment change. The assortment change includes a set of one or more items to be added to a current item assortment and/or a set of one or more items to be removed from the current item assortment. The demand prediction component generates a demand transference result including the calculated magnitude of demand transfer for each item in the set of substitute items and/or a predicted walk-off rate associated with lost demand. An assortment recommendation component generates an accept recommendation and/or a reject recommendation based on the demand transference result, the predicted walk-off rate, and/or a demand transference score.

Abstract: In some embodiments, methods and systems of detecting and monitoring shoplifting of products by individuals at a retail facility include sensors coupled to a product and located at a fixed threshold distance relative to each other. The sensors are configured to generate an alert signal indicating that the fixed threshold distance between the first and second sensors has been altered. An alert sensor positioned in the product stocking area of the retail facility is configured to receive the alert signal from one or both of the product sensors. A central computing device at the retail facility is configured to determine the alert sensor that received the alert signal, and to activate a video camera in order to acquire a photo or video of the portion of the product stocking area of the retail facility associated with the alert sensor that received the alert signal.

Abstract: In some embodiments, methods and systems of detecting and monitoring shoplifting of products by individuals at a retail facility include sensors coupled to a product and located at a fixed threshold distance relative to each other. The sensors are configured to generate an alert signal indicating that the fixed threshold distance between the first and second sensors has been altered. An alert sensor positioned in the product stocking area of the retail facility is configured to receive the alert signal from one or both of the product sensors. A central computing device at the retail facility is configured to determine the alert sensor that received the alert signal, and to activate a video camera in order to acquire a photo or video of the portion of the product stocking area of the retail facility associated with the alert sensor that received the alert signal.