Trend Detection in a Messaging Platform
AssigneeTwitter, Inc.
InventorsScott Hendrickson, Jeffrey Aaron Kolb, Joshua Montague, Brian Lehman
Life of this patent
Estimating the term…
Abstract
A system and method for trend detection in a messaging platform. A trend detection model is selected and a time series having a plurality of instances of social data is received, wherein the instances of social data share a countable parameter. A count is made of occurrences of countable parameters in each instance of social data assigned to that bin and a trend detected based at least in part on the trend detection model and on the count for each bin.
Claims18
A method for quantifying changes in social data and accounting for statistical false positives in the quantified changes in the social data so as to detect one or more trends in the social data, the method comprising:
- labeling, by a computing device, a set of historical time series of social data, wherein each historical time series of the set of historical time series is labeled by the computing device as trending and non-trending;
- selecting, by the computing device, based at least in part on the labeled set of historical time series of social data, a trend detection model;
- receiving, by the computing device, a sampled time series of social data having a plurality of instances of social data, wherein one or more instances of the social data in the sampled time series of social data correspond to a particular event;
- determining, by the computing device using the trend detection model, a distance between at least one labeled historical time series of social data from the labeled set of historical times series of the social data and the sampled time series of the social data;
- determining a measure of the trend based at least in part on the distance determined using the trend detection model;
- implementing cycle-correction to the measure of the trend, based on patterns associated with respective time data associated with each historical time series of the set of historical time series labeled as trending, to obtain a cycle-corrected measure of trend;
- determining, by the computing device, that the cycle-corrected measure of the trend satisfies a trend threshold;
- and
- responsive to determining that the cycle-corrected measure of the trend satisfies the trend threshold, outputting, by the computing device, at least one indication of the detected trend.
A computing device for quantifying changes in social media and accounting for statistical false positives in the quantified changes in the social media so as to detect one or more trends in the social media, the computing device comprising:
- at least one processor;
- and
- at least one non-transitory computer-readable storage medium storing instructions that are executable by the at least one processor to:
- label a set of historical time series of social data, wherein each historical time series of the set of historical time series is labeled as trending and non-trending;
- select based at least in part on the labeled set of historical time series of social data, a trend detection model;
- receive a sampled time series of social data having a plurality of instances of social data, wherein one or more instances of the social data in the sampled time series of social data correspond to a particular event;
- determine using the trend detection model a distance between at least one labeled historical time series of social data from the labeled set of historical time series of social data and the sampled time series of the social data;
- determine a measure of the trend based at least in part on the distance determined using the trend detection model;
- implement cycle-correction to the measure of the trend, based on patterns associated with respective time data associated with each historical time series of the set of historical time series labeled as trending, to obtain a cycle-corrected measure of trend;
- determine that the cycle-corrected measure of the trend satisfies a trend threshold;
- and
- responsive to determining that the cycle-corrected measure of the trend satisfies the trend threshold, output at least one indication of the detected trend.
A non-transitory computer-readable storage medium for quantifying changes in social data and accounting for statistical false positives in the quantified changes in the social data so as to detect one or more trends in the social data encoded with instructions that, when executed, cause at least one processor of a computing device to:
- label a set of historical time series of social data, wherein each historical time series of the set of historical time series is labeled as trending and non-trending;
- select based at least in part on the labeled set of historical time series of social data, a trend detection model;
- receive a sampled time series of social data having a plurality of instances of social data, wherein one or more instances of the social data in the sampled time series of social data correspond to a particular event;
- determine using a trend detection model, a distance between at least one labeled historical time series of social data from the labeled set of historical time series of social data and the sampled time series of the social data;
- determine a measure of the trend based at least in part on the distance determined using the trend detection model;
- implement cycle-correction to the measure of the trend, based on patterns associated with respective time data associated with each historical time series of the set of historical time series labeled as trending, to obtain a cycle-corrected measure of trend;
- determine that the cycle-corrected measure of the trend satisfies a trend threshold;
- and
- responsive to determining that the cycle-corrected measure of the trend satisfies the trend threshold, output at least one indication of the detected trend.
Dependencies are inferred from the claim text.