Automated alarm panel classification using pareto optimization

Abstract

Systems and methods for alarm panel analysis include receiving a plurality of alarm events from the respective alarm panels monitoring corresponding buildings. An alarm analysis system may classify the alarm panels by identifying an alarm type for the alarm events. The alarm analysis system may determine a number of occurrences of each alarm type for the alarm panels, and generate a data point for a dataset for the alarm panels. The alarm analysis system may identify statistical dividers in the data points for the data set, which may be used for assigning a ranking for the alarm panels based on their location in relation to the statistical dividers. The alarm analysis system may construct and a monitoring dashboard which includes the rankings of the alarm panels, which may be rendered on a display to an end user.

What is claimed is: 
     
       1. A system for analyzing alarm panels, the system comprising:
 a communications interface communicably coupled to a plurality of alarm panels at respective buildings, the communications interface configured to receive alarm events from the alarm panels, the alarm events indicating an alarm type; and 
 a processing circuit configured to:
 receive, via the communications interface, a plurality of alarm events from the respective alarm panels; 
 classify each of the alarm panels according to the plurality of alarm events by:
 identifying, for the alarm events, an alarm type; 
 determining, for the alarm panels, a number of occurrences of each alarm type; 
 generating, for the alarm panels, a data point for a dataset, the data point representing the number of occurrences of the alarm types for a respective alarm panel; 
 identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset; and 
 assign a ranking for the alarm panels based on their respective location in relation to the statistical dividers; 
 
 construct a monitoring dashboard which includes the ranking of the alarm panels according to the classification of the alarm panels; and 
 cause the monitoring dashboard to be rendered on a display to an end user, the monitoring dashboard indicating the ranking of the alarm panels. 
 
 
     
     
       2. The system of  claim 1 , wherein the statistical dividers are parallel and equidistant from one another. 
     
     
       3. The system of  claim 1 , wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers; 
 computing an upper Pareto frontier within the dataset; 
 computing a lower Pareto frontier within the dataset; 
 computing an upper centroid for the upper Pareto frontier, a lower centroid for the lower Pareto frontier, and a midpoint for the upper centroid and lower centroid; and 
 defining the statistical dividers including an upper statistical divider, a lower statistical divider, and a middle statistical divider in relation to at least one of the upper centroid, lower centroid, and midpoint. 
 
     
     
       4. The system of  claim 3 , wherein computing the upper centroid, the lower centroid, and the midpoint comprises:
 computing the upper centroid, C UPF , for the upper Pareto frontier; 
 computing the lower centroid, C LPF , for the lower Pareto frontier; 
 computing the midpoint, C MPF , between the upper centroid and lower centroid according to C MPF =(C UPF +C LPF )/2. 
 
     
     
       5. The system of  claim 4 , wherein identifying the statistical dividers comprises:
 selecting a value, α, between 0 and 0.5 which defines a spacing between the statistical dividers; 
 defining an upper statistical divider centroid, C VH , as C VH =(1+2α)C MPF ; 
 defining a lower statistical divider centroid, C VL , as C VL =(1−2α)C MPF ; 
 defining the middle statistical divider, which bisects the midpoint; 
 defining the upper statistical divider and lower statistical divider as extending through the upper statistical divider centroid and lower statistical centroid and extending parallel to the middle statistical divider. 
 
     
     
       6. The system of  claim 5 , wherein identifying the statistical dividers further comprises:
 defining a middle upper statistical divider centroid, C H , as C H =(1+α)C MPF ; and 
 defining a middle lower statistical divider centroid, C L , as C L =(1−α)C MPF . 
 
     
     
       7. The system of  claim 6 , wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and separator 
 assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location. 
 
     
     
       8. The system of  claim 1 , wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers to generate a filtered dataset D; 
 computing an upper Pareto frontier (UPF D ) for the filtered dataset D; 
 selecting a value, α, between 0 and 0.25 which defines a spacing between the statistical dividers; and 
 computing a plurality of lower Pareto frontiers including a very high Pareto frontier (UPF VH ), a high Pareto frontier (UPF H ), a middle Pareto frontier (UPF M ), a low Pareto frontier (UPF L ), and a very low Pareto frontier (UPF VL ) according to:
   UPF VH =UPF D    
   UPF H =(1−α)UPF D  
 
   UPF M =(1−2α)UPF D  
 
   UPF L =(1−3α)UPF D  
 
   UPF VL =(1−4α)UPF D ;
 
 
 wherein the upper statistical divider is UPF VH , the lower statistical divider is UPF VL , the middle statistical divider is UPF M , and wherein define the statistical dividers further include an upper middle statistical divider UPF H  and a lower middle statistical divider UPF L . 
 
     
     
       9. The system of  claim 8 , wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and 
 assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location. 
 
     
     
       10. The system of  claim 8 , wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, which statistical divider a data point is located between from the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and 
 assigning a ranking of very high, high, middle, low, and very low to the data point based on which statistical divider the data point is located between. 
 
     
     
       11. A method for analyzing alarm panels, the system comprising:
 receiving, via a communications interface communicably coupled to a plurality of alarm panels at respective buildings, a plurality of alarm events from the respective alarm panels, wherein the communications interface is configured to receive alarm events from the alarm panels, the alarm events indicating an alarm type; 
 classifying the alarm panels according to the plurality of alarm events by:
 identifying, for the alarm events, an alarm type; 
 determining, for the alarm panels, a number of occurrences of each alarm type; 
 generating, for the alarm panels, a data point for a dataset, the data point representing the number of occurrences of the alarm types for a respective alarm panel; 
 identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset; and 
 assigning a ranking for the alarm panels based on their respective location in relation to the statistical dividers; 
 
 constructing a monitoring dashboard which includes the ranking of the alarm panels according to the classification of the alarm panels; and 
 causing the monitoring dashboard to be rendered on a display to an end user, the monitoring dashboard indicating the ranking of the alarm panels. 
 
     
     
       12. The method of  claim 11 , wherein the statistical dividers are parallel and equidistant from one another. 
     
     
       13. The method of  claim 11 , wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers; 
 computing an upper Pareto frontier within the dataset; 
 computing a lower Pareto frontier within the dataset; 
 computing an upper centroid for the upper Pareto frontier, a lower centroid for the lower Pareto frontier, and a midpoint for the upper centroid and lower centroid; and 
 defining the statistical dividers including an upper statistical divider, a lower statistical divider, and a middle statistical divider in relation to at least one of the upper centroid, lower centroid, and midpoint. 
 
     
     
       14. The method of  claim 13 , wherein computing the upper centroid, the lower centroid, and the midpoint comprises:
 computing the upper centroid, C UPF , for the upper Pareto frontier; 
 computing the lower centroid, C LPF , for the lower Pareto frontier; 
 computing the midpoint, C MPF , between the upper centroid and lower centroid according to C MPF =(C UPF +C LPF )/2. 
 
     
     
       15. The method of  claim 14 , wherein identifying the statistical dividers comprises:
 selecting a value, α, between 0 and 0.5 which defines a spacing between the statistical dividers; 
 defining an upper statistical divider centroid, C VH , as C VH =(1+2α)C MPF ; 
 defining a middle upper statistical divider centroid, C H , as C H =(1+α)C MPF ; 
 defining a middle lower statistical divider centroid, C L , as C L =(1−α)C MPF ; 
 defining a lower statistical divider centroid, C VL , as C VL =(1−2α)C MPF ; 
 defining the middle statistical divider, which bisects the midpoint; and 
 defining the upper statistical divider, a middle upper statistical divider, a middle lower statistical divider, and the lower statistical divider, which extend through the upper statistical divider centroid, the middle upper statistical divider centroid, the middle lower statistical divider centroid, and the lower statistical divider centroid, and parallel to the middle statistical divider. 
 
     
     
       16. The method of  claim 15 , wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider, each data point; and 
 assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location. 
 
     
     
       17. The method of  claim 11 , wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers to generate a filtered dataset D; 
 computing an upper Pareto frontier (UPF D ) for the filtered dataset D; 
 selecting a value, α, between 0 and 0.25 which defines a spacing between the statistical dividers; and 
 computing a plurality of lower Pareto frontiers including a very high Pareto frontier (UPF VH ), a high Pareto frontier (UPF H ), a middle Pareto frontier (UPF M ), a low Pareto frontier (UPF L ), and a very low Pareto frontier (UPF VL ) according to:
   UPF VH =UPF D    
   UPF H =(1−α)UPF D  
 
   UPF M =(1−2α)UPF D  
 
   UPF L =(1−3α)UPF D  
 
   UPF VL =(1−4α)UPF D ;
 
 
 wherein the upper statistical divider is UPF VH , the lower statistical divider is UPF VL , the middle statistical divider is UPF M , and wherein define the statistical dividers further include an upper middle statistical divider UPF H  and a lower middle statistical divider UPF L . 
 
     
     
       18. The method of  claim 17 , wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and 
 assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location. 
 
     
     
       19. The method of  claim 17 , wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, which statistical divider a data point is located between from the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and 
 assigning a ranking of very high, high, middle, low, and very low to the data point based on which statistical divider the data point is located between. 
 
     
     
       20. A computing device for analyzing alarm panels, the computing device comprising:
 a processing circuit including a processor and memory, the memory storing instructions that, when executed by the processor, cause the processor to:
 receive, via a communications interface communicably coupled to a plurality of alarm panels at respective buildings, a plurality of alarm events from the respective alarm panels, wherein the communications interface is configured to receive alarm events from the alarm panels, the alarm events indicating an alarm type; 
 classify the alarm panels according to the plurality of alarm events by: 
 identifying, for the alarm events, an alarm type; 
 determining, for the alarm panels, a number of occurrences of each alarm type; 
 generating, for the alarm panels, a data point for a dataset, the data point representing the number of occurrences of the alarm types for a respective alarm panel; 
 identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset; and assigning a ranking for the alarm panels based on their respective location in relation to the statistical dividers; 
 
 construct a monitoring dashboard which includes the ranking of the alarm panels according to the classification of the alarm panels; and 
 cause the monitoring dashboard to be rendered on a display to an end user, the monitoring dashboard. 
 
     
     
       21. A system for analyzing alarm panels, the system comprising:
 a communications interface communicably coupled to a plurality of alarm panels at respective buildings, the communications interface configured to receive alarm events from the alarm panels, the alarm events indicating an alarm type; and   a processing circuit configured to:
 receive, via the communications interface, a plurality of alarm events from the respective alarm panels; 
 classify each of the alarm panels according to the plurality of alarm events by:
 identifying, for the alarm events, an alarm type; 
 determining, for the alarm panels, a number of occurrences of each alarm type; 
 generating, for the alarm panels, a data point for a dataset, the data point representing the number of occurrences of the alarm types for a respective alarm panel; and 
 assign a ranking for the alarm panels based on the respective data points within the dataset; 
 
 construct a monitoring dashboard which includes the ranking of the alarm panels according to the classification of the alarm panels; and 
 cause the monitoring dashboard to be rendered on a display to an end user, the monitoring dashboard indicating the ranking of the alarm panels. 
   
     
     
       22. The system of claim 21, wherein the processing circuit is further configured to classify each of the alarm panels according to the plurality of alarm events by identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset, wherein the statistical dividers are parallel and equidistant from one another. 
     
     
       23. The system of claim 21, wherein the processing circuit is further configured to classify each of the alarm panels according to the plurality of alarm events by identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset, wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers;   computing an upper Pareto frontier within the dataset;   computing a lower Pareto frontier within the dataset;   computing an upper centroid for the upper Pareto frontier, a lower centroid for the lower Pareto frontier, and a midpoint for the upper centroid and lower centroid; and   defining the statistical dividers including an upper statistical divider, a lower statistical divider, and a middle statistical divider in relation to at least one of the upper centroid, lower centroid, and midpoint.   
     
     
       24. The system of claim 23, wherein computing the upper centroid, the lower centroid, and the midpoint comprises:
 computing the upper centroid, C UPF , for the upper Pareto frontier;   computing the lower centroid, C LPF , for the lower Pareto frontier;   computing the midpoint, C MPF , between the upper centroid and lower centroid according to C MPF =(C UPF +C LPF )/2.   
     
     
       25. The system of claim 24, wherein identifying the statistical dividers comprises:
 selecting a value, α, between 0 and 0.5 which defines a spacing between the statistical dividers;   defining an upper statistical divider centroid, C VH , as C VH =(1+2α)C MPF ;   defining a lower statistical divider centroid, C VL , as C VL =(1−2α)C MPF ;   defining the middle statistical divider, which bisects the midpoint;   defining the upper statistical divider and lower statistical divider as extending through the upper statistical divider centroid and lower statistical centroid and extending parallel to the middle statistical divider.   
     
     
       26. The system of claim 25, wherein identifying the statistical dividers further comprises:
 defining a middle upper statistical divider centroid, C H , as C H =(1+α)C MPF ; and   defining a middle lower statistical divider centroid, C L , as C L =(1−α)C MPF .   
     
     
       27. The system of claim 26, wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and separator   assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location.   
     
     
       28. The system of claim 21, wherein the processing circuit is further configured to classify each of the alarm panels according to the plurality of alarm events by identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset, wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers to generate a filtered dataset D;   computing an upper Pareto frontier (UPF D ) for the filtered dataset D;   selecting a value, α, between 0 and 0.25 which defines a spacing between the statistical dividers; and   computing a plurality of lower Pareto frontiers including a very high Pareto frontier (UPF VH ), a high Pareto frontier (UPF H ), a middle Pareto frontier (UPF M ), a low Pareto frontier (UPF L ), and a very low Pareto frontier (UPF VL ) according to:
   UPF VH =UPF D    
   UPF H =(1−α)UPF D  
 
   UPF M =(1−2α)UPF D  
 
   UPF L =(1−3α)UPF D  
 
   UPF VL =(1−4α)UPF D ;
 
   wherein the upper statistical divider is UPF VH , the lower statistical divider is UPF VL , the middle statistical divider is UPF M , and wherein define the statistical dividers further include an upper middle statistical divider UPF H  and a lower middle statistical divider UPF L .   
     
     
       29. The system of claim 28, wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and   assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location.   
     
     
       30. The system of claim 28, wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, which statistical divider a data point is located between from the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and   assigning a ranking of very high, high, middle, low, and very low to the data point based on which statistical divider the data point is located between.   
     
     
       31. A method for analyzing alarm panels, the system comprising:
 receiving, via a communications interface communicably coupled to a plurality of alarm panels at respective buildings, a plurality of alarm events from the respective alarm panels, wherein the communications interface is configured to receive alarm events from the alarm panels, the alarm events indicating an alarm type;   classifying the alarm panels according to the plurality of alarm events by:
 identifying, for the alarm events, an alarm type; 
 determining, for the alarm panels, a number of occurrences of each alarm type; 
 generating, for the alarm panels, a data point for a dataset, the data point representing the number of occurrences of the alarm types for a respective alarm panel; and 
 assigning a ranking for the alarm panels based on the respective data points within the dataset; 
   constructing a monitoring dashboard which includes the ranking of the alarm panels according to the classification of the alarm panels; and   causing the monitoring dashboard to be rendered on a display to an end user, the monitoring dashboard indicating the ranking of the alarm panels.   
     
     
       32. The method of claim 31, wherein classifying the alarm panels according to the plurality of alarm events further comprises identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset, wherein the statistical dividers are parallel and equidistant from one another. 
     
     
       33. The method of claim 31, wherein classifying the alarm panels according to the plurality of alarm events further comprises identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset, wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers;   computing an upper Pareto frontier within the dataset;   computing a lower Pareto frontier within the dataset;   computing an upper centroid for the upper Pareto frontier, a lower centroid for the lower Pareto frontier, and a midpoint for the upper centroid and lower centroid; and   defining the statistical dividers including an upper statistical divider, a lower statistical divider, and a middle statistical divider in relation to at least one of the upper centroid, lower centroid, and midpoint.   
     
     
       34. The method of claim 33, wherein computing the upper centroid, the lower centroid, and the midpoint comprises:
 computing the upper centroid, C UPF , for the upper Pareto frontier;   computing the lower centroid, C LPF , for the lower Pareto frontier;   computing the midpoint, C MPF , between the upper centroid and lower centroid according to C MPF =(C UPF +C LPF )/2.   
     
     
       35. The method of claim 34, wherein identifying the statistical dividers comprises:
 selecting a value, α, between 0 and 0.5 which defines a spacing between the statistical dividers;   defining an upper statistical divider centroid, C VH , as C VH =(1+2α)C MPF ;   defining a middle upper statistical divider centroid, C H , as C H =(1+α)C MPF ;   defining a middle lower statistical divider centroid, C L , as C L =(1−α)C MPF ;   defining a lower statistical divider centroid, C VL , as C VL =(1−2α)C MPF ;   defining the middle statistical divider, which bisects the midpoint; and   defining the upper statistical divider, a middle upper statistical divider, a middle lower statistical divider, and the lower statistical divider, which extend through the upper statistical divider centroid, the middle upper statistical divider centroid, the middle lower statistical divider centroid, and the lower statistical divider centroid, and parallel to the middle statistical divider.   
     
     
       36. The method of claim 35, wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider, each data point; and   assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location.   
     
     
       37. The method of claim 31, wherein classifying the alarm panels according to the plurality of alarm events further comprises identifying, for the dataset, statistical dividers in the data points, the statistical dividers defining a separation of rankings for the data points within the dataset, wherein identifying the statistical dividers in the data points comprises:
 applying a filter to the data points of the dataset which removes zeros and outliers to generate a filtered dataset D;   computing an upper Pareto frontier (UPF D ) for the filtered dataset D;   selecting a value, α, between 0 and 0.25 which defines a spacing between the statistical dividers; and   computing a plurality of lower Pareto frontiers including a very high Pareto frontier (UPF VH ), a high Pareto frontier (UPF H ), a middle Pareto frontier (UPF M ), a low Pareto frontier (UPF L ), and a very low Pareto frontier (UPF VL ) according to:
   UPF VH =UPF D    
   UPF H =(1−α)UPF D  
 
   UPF M =(1−2α)UPF D  
 
   UPF L =(1−3α)UPF D  
 
   UPF VL =(1−4α)UPF D ;
 
   wherein the upper statistical divider is UPF VH , the lower statistical divider is UPF VL , the middle statistical divider is UPF M , and wherein define the statistical dividers further include an upper middle statistical divider UPF H  and a lower middle statistical divider UPF L .   
     
     
       38. The method of claim 37, wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, a relative location to the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and   assigning a ranking of very high, high, middle, low, and very low to each data point according to the relative location.   
     
     
       39. The method of claim 37, wherein assigning the ranking comprises:
 determining, for each data point in the dataset D, which statistical divider a data point is located between from the upper statistical divider, upper middle statistical divider, middle statistical divider, lower middle statistical divider, and lower statistical divider; and   assigning a ranking of very high, high, middle, low, and very low to the data point based on which statistical divider the data point is located between.   
     
     
       40. A computing device for analyzing alarm panels, the computing device comprising:
 a processing circuit including a processor and memory, the memory storing instructions that, when executed by the processor, cause the processor to:
 receive, via a communications interface communicably coupled to a plurality of alarm panels at respective buildings, a plurality of alarm events from the respective alarm panels, wherein the communications interface is configured to receive alarm events from the alarm panels, the alarm events indicating an alarm type; 
 classify the alarm panels according to the plurality of alarm events by: 
 identifying, for the alarm events, an alarm type; 
 determining, for the alarm panels, a number of occurrences of each alarm type; 
 generating, for the alarm panels, a data point for a dataset, the data point representing the number of occurrences of the alarm types for a respective alarm panel; and 
 assigning a ranking for the alarm panels based on the respective data points within the dataset; 
   construct a monitoring dashboard which includes the ranking of the alarm panels according to the classification of the alarm panels; and   cause the monitoring dashboard to be rendered on a display to an end user, the monitoring dashboard.