Clari Inc.

Walmart Apollo, LLC

Lei Tang

Mahesh Subedi

Kurt Leafstrand

MohamadAli Torkamani

Xin Xu

Venkat Rangan

Patrick Harrington

Tao Zhu

Mark Shuai Sun

Activity data of a set of tasks as a training set is obtained from a list of communication platforms associated with the tasks. For each of the tasks in the training set, a set of activity metrics is compiled according to a set of predetermined activity categories based on the activity data of each task. The activity metrics of all of the tasks in the training set are aggregated based on the activity categories to generate a data matrix. A principal component analysis is performed on the metrics of its covariance matrix to derive an activity dimension vector, where the activity dimension vector represents a distribution pattern of the activity metrics of the tasks. The activity dimension vector can be utilized to determine an activity score of a particular task, where the activity score of a task can be utilized to estimate a probability of completeness of the task.

Method and system for summarizing user activities of tasks into a single activity score using machine learning to predict probabilities of completeness of the tasks

A request is received for determining a task completion rate of each of a first set of tasks associated with a set of task attributes. The first set of tasks are scheduled to be completed within a first timer period. An MAPE score is calculated or obtained for each of the completion rate predictive models, which is determined based on prior predictions performed in a second time period in the past. The duration of the second time period is a multiple of the first time period. One of the predictive models is selected based on the MAPE scores of the predictive models, where the selected model has the lowest MAPE score amongst the predictive models in the set. In another embodiment, a predictive model is selected further based on the volatility scores of the predictive models. A model with a combination of lowest MAPE score and volatility score is selected.

Method and system for predicting task completion of a time period based on task completion rates and data trend of prior time periods in view of attributes of tasks using machine learning models

Providing personalized item recommendations using scalable matrix factorization with randomness. Various embodiments involve ecommerce websites, an item interaction matrix, item interaction weightings, factorizing into a user feature matrix and an item feature matrix, computing a thin matrix, generating a random Gaussian matrix, decomposing a matrix Y by a QR factorization, performing a singular value decomposition, performing alternative least squares, computing item interaction scores, selecting at least one item for recommendation to an Internet user based on the item interaction scores, or a combination thereof.

Providing personalized item recommendations using scalable matrix factorization with randomness

A request is received from a client for determining task completion of a first set of tasks associated with attributes, the first set of tasks scheduled to be performed within a first time period. For each of the attributes, a completion rate of one or more of a second set of tasks is calculated that are associated with the attribute. The second set of tasks has been performed during a second time period in the past. An isotonic regression operation and/or temporal smoothing are performed on the completion rates associated with the attributes of the second set of tasks that have been performed during the second time period to calibrate the completion rates. Possible completion for the attributes of the first set of tasks to be performed in the first time period is calculated based on the calibrated completion rates of the second set of tasks.

Growing community of inventors

Milpitas, CA, United States of America

Lei Tang