Abstract: According to an embodiment of the present invention, an approach accurately detects anomalies or outliers of a time-series dataset. A method for identifying whether a particular data element of the time-series dataset is an outlier comprises predicting a value for that particular data element and obtaining a threshold value that defines, relative to the predicted value, whether an actual value of the data element is an outlier. In an aspect of a present invention embodiment, the threshold value is generated based on historic error values associated with data elements temporally preceding the particular data element of the time-series dataset.

Abstract: An embodiment for contextualizing abnormal events which employ location-based social networks, LBSN, data to determine events that may be linked to the abnormal events is provided. The embodiment may include detecting an occurrence of an abnormal event within a geographic region, wherein the abnormal event occurs at an occurrence time. The embodiment may also include obtaining location-based social networks, LBSN, data relating to the geographic region for a time period including the occurrence time. The embodiment may further include analyzing the obtained LBSN data, wherein the analyzation determines a linked event within the geographic region for the time period. The embodiment may also include associating the linked event with the abnormal event.

Abstract: An embodiment for contextualizing abnormal events which employ location-based social networks, LBSN, data to determine events that may be linked to the abnormal events is provided. The embodiment may include detecting an occurrence of an abnormal event within a geographic region, wherein the abnormal event occurs at an occurrence time. The embodiment may also include obtaining location-based social networks, LBSN, data relating to the geographic region for a time period including the occurrence time. The embodiment may further include analyzing the obtained LBSN data, wherein the analyzation determines a linked event within the geographic region for the time period. The embodiment may also include associating the linked event with the abnormal event.

Abstract: In an approach to managing transmission frequency for a plurality of edge devices of an interconnected distributed network, one or more computer processors determine a maximum writing frequency, MWF, for the interconnected distributed network; iteratively process values of data flow writing frequency, DFWF, for a plurality of edge devices of an interconnected distributed network in accordance with an optimization algorithm based on the MWF to identify a convergence in the values of DFWF, wherein each iteration of processing values comprises, at each edge device in the plurality of edge devices, determine a value of DFWF based on an associated utility function of the respective edge device, wherein the utility function is a measure of utility of the device as a function of DFWF; responsive to identifying convergence, determine the converged values of DFWF to be optimal values of DFWF for the plurality of edge devices.

Abstract: In an approach to managing transmission frequency for a plurality of edge devices of an interconnected distributed network, one or more computer processors determine a maximum writing frequency, MWF, for the interconnected distributed network; iteratively process values of data flow writing frequency, DFWF, for a plurality of edge devices of an interconnected distributed network in accordance with an optimization algorithm based on the MWF to identify a convergence in the values of DFWF, wherein each iteration of processing values comprises, at each edge device in the plurality of edge devices, determine a value of DFWF based on an associated utility function of the respective edge device, wherein the utility function is a measure of utility of the device as a function of DFWF; responsive to identifying convergence, determine the converged values of DFWF to be optimal values of DFWF for the plurality of edge devices.

Abstract: A method, computer system, and a computer program product for improving social media network actions is provided. The present invention may include determining a proposed action by a user and predicting a predicted influence score for the proposed action. The present invention may include identifying a prior related action in the user's social media network. The present invention may include retrieving a previously calculated average influence score for the type and user of the prior related action and applying a decay profile to the average influence score from the time of the action to a current time. The present invention may include comparing the predicted influence score with the decayed average influence score at the current time. The present invention may include posting the proposed action on the social media network at a time when the predicted influence score is greater than the decayed average influence score.

Abstract: Detecting data anomalies by receiving a first data set related to a first variable metric, determining data anomaly detection scores for data points of the first data set according to a plurality of data anomaly detection techniques, generating an adaptive ground-truth window according to the data anomaly detection scores, assigning a weighting value to each data point within the adaptive ground-truth window, training a machine learning system using the set of data anomaly detection scores and weighting values, and providing a trained machine learning system for evaluating a second data set.

Abstract: Detecting data anomalies by receiving a first data set related to a first variable metric, determining data anomaly detection scores for data points of the first data set according to a plurality of data anomaly detection techniques, generating an adaptive ground-truth window according to the data anomaly detection scores, assigning a weighting value to each data point within the adaptive ground-truth window, training a machine learning system using the set of data anomaly detection scores and weighting values, and providing a trained machine learning system for evaluating a second data set.

Abstract: A method, computer system, and a computer program product for improving social media network actions is provided. The present invention may include determining a proposed action by a user and predicting a predicted influence score for the proposed action. The present invention may include identifying a prior related action in the user's social media network. The present invention may include retrieving a previously calculated average influence score for the type and user of the prior related action and applying a decay profile to the average influence score from the time of the action to a current time. The present invention may include comparing the predicted influence score with the decayed average influence score at the current time. The present invention may include posting the proposed action on the social media network at a time when the predicted influence score is greater than the decayed average influence score.

Abstract: Various embodiments are directed to concepts for spatio-temporal prediction based on one-dimensional features and two-dimensional features from diverse data sources. One embodiment comprises processing one-dimensional data matrices representative of variations of one-dimensional, 1D, feature values with a fully connected network to generate respective outputs from the fully connected network. It also comprises processing two-dimensional data matrices representative of variations of two-dimensional, 2D, feature values with a convolutional neural network to generate respective outputs from the convolutional neural network.

Abstract: According to an embodiment of the present invention, an approach accurately detects anomalies or outliers of a time-series dataset. A method for identifying whether a particular data element of the time-series dataset is an outlier comprises predicting a value for that particular data element and obtaining a threshold value that defines, relative to the predicted value, whether an actual value of the data element is an outlier. In an aspect of a present invention embodiment, the threshold value is generated based on historic error values associated with data elements temporally preceding the particular data element of the time-series dataset.