System, method, and computer program product for generating error rate predictions based on machine learning using incremental backpropagation

1 . A system comprising:

at least one processor programmed or configured to:

receive a dataset of a plurality of data instances, wherein the plurality of data instances comprises at least one data instance corresponding to a first category of parameters, at least one data instance corresponding to a second category of parameters, and at least one data instance corresponding to a third category of parameters;

train a first prediction model based on the data instances corresponding to the first category of parameters, wherein the first category of parameters comprises attributes, wherein an attribute is a text string and the at least one data instance corresponding to the first category of parameters comprises a plurality of data points, wherein each data point is an integer representing the attribute that is to be included or not included in a data report, wherein the first prediction model is configured to provide an output that comprises a first predicted error rate, and wherein the first predicted error rate represents a probability that generating the data report will produce an error based on a parameter in the first category of parameters that is to be included or not included in the data report;

train a second prediction model based on the data instances corresponding to the second category of parameters, wherein the second category of parameters comprises metrics, wherein a metric is a numerical value and the at least one data instance corresponding to the second category of parameters comprises a plurality of data points, wherein each data point is an integer representing the metric that is to be included or not included in a data report, wherein the second prediction model is configured to provide an output that comprises a second predicted error rate, and wherein the second predicted error rate represents a probability that generating the data report will produce an error based on a parameter in the second category of parameters that is to be included or not included in the data report;

train a third prediction model based on the data instances corresponding to the third category of parameters, wherein the third category of parameters comprises filters, wherein a filter is a predefined text string or a predefined numerical value, and the at least one data instance corresponding to the third category of parameters comprises a plurality of data points, wherein each data point is an integer equal to 1 or 0, wherein the integer equal to 1 represents the filter that is included in a data report and the integer equal to 0 represents the filter that is not included in the data report, wherein the third prediction model is configured to provide an output that comprises a third predicted error rate, and wherein the third predicted error rate represents a probability that generating a data report will produce an error based on a parameter in the third category of parameters that is to be included or not included in the data report;

train an ensemble machine learning model based on an output of the first prediction model, an output of the second prediction model, and an output of the third prediction model to provide a trained ensemble machine learning model; and

generate a runtime output of the trained ensemble machine learning model based on a runtime input, wherein the runtime output comprises a prediction of a total error rate and an indication of a predicted cause of the total error rate based on one or more parameters selected by a user, and wherein, when generating the runtime output of the trained ensemble machine learning model, the at least one processor is programmed or configured to:

generate a first runtime error rate with the first prediction model;

generate a second runtime error rate with the second prediction model;

generate a third runtime error rate with the third prediction model;

generate a total error rate based on the first runtime error rate, the second runtime error rate, and the third runtime error rate; and

generate an indication of a predicted cause of the total error rate with the trained ensemble machine learning model based on the one or more parameters selected by the user for a data report associated with the runtime input.

2 . The system of claim 1 , wherein the at least one processor is further programmed or configured to:

perform an action based on the runtime output of the trained ensemble machine learning model.

3 . The system of claim 1 , wherein, when receiving the dataset of the plurality of data instances, the at least one processor is programmed or configured to:

perform a normalization operation on the plurality of data instances to provide a plurality of normalized data instances, wherein each data instance comprises a plurality of data points, and wherein each data point has a value from 0 to 1.

4 . The system of claim 1 , wherein, when generating a runtime output of the trained ensemble machine learning model, the at least one processor is programmed or configured to:

receive a first data instance corresponding to the first category of parameters, wherein the first data instance comprises a plurality of data points, and wherein each data point is an integer representing a first parameter that is selected by the user to be included or not included in the data report;

receive a second data instance corresponding to the second category of parameters, wherein the second data instance comprises a plurality of data points, and wherein each data point is an integer representing a second parameter that is selected by the user to be included or not included in the data report;

receive a third data instance corresponding to the third category of parameters, wherein the third data instance comprises a plurality of data points, and wherein each data point is an integer representing a third parameter that is selected by the user to be included or not included in the data report;

generate a first runtime error rate with the first prediction model based on the first data instance;

generate a second runtime error rate with the second prediction model based on the second data instance;

generate a third runtime error rate with the third prediction model based on the third data instance; and

generate the total error rate and the indication of the predicted cause of the total error rate with the trained ensemble machine learning model based on the first runtime error rate, the second runtime error rate, and the third runtime error rate.

5 . The system of claim 1 , wherein the first category of parameters is attributes, wherein an attribute is a text string and the at least one data instance corresponding to the first category of parameters comprises a plurality of data points, and wherein each data point is an integer representing the attribute that is to be included or not included in the data report.

6 . The system of claim 1 , wherein the second category of parameters is metrics, wherein a metric is a numerical value and the at least one data instance corresponding to the second category of parameters comprises a plurality of data points, and wherein each data point is an integer representing the metric that is to be included or not included in the data report.

7 . The system of claim 1 , wherein the third category of parameters is filters, wherein a filter is a predefined text string or a predefined numerical value, and the at least one data instance corresponding to the third category of parameters comprises a plurality of data points, wherein each data point is an integer equal to 1 or 0, and wherein the integer equal to 1 represents the filter that is included in the data report and the integer equal to 0 represents the filter that is not included in the data report.

8 . A method comprising:

receiving, with at least one processor, a dataset of a plurality of data instances, wherein the plurality of data instances comprises at least one data instance corresponding to a first category of parameters, at least one data instance corresponding to a second category of parameters, and at least one data instance corresponding to a third category of parameters;

training, with the at least one processor, a first prediction model based on the data instances corresponding to the first category of parameters, wherein the first category of parameters comprises attributes, wherein an attribute is a text string and the at least one data instance corresponding to the first category of parameters comprises a plurality of data points, wherein each data point is an integer representing the attribute that is to be included or not included in a data report, wherein the first prediction model is configured to provide an output that comprises a first predicted error rate, and wherein the first predicted error rate represents a probability that generating the data report will produce an error based on a parameter in the first category of parameters that is to be included or not included in the data report;

training, with the at least one processor, a second prediction model based on the data instances corresponding to the second category of parameters, wherein the second category of parameters comprises metrics, wherein a metric is a numerical value and the at least one data instance corresponding to the second category of parameters comprises a plurality of data points, wherein each data point is an integer representing the metric that is to be included or not included in the data report, wherein the second prediction model is configured to provide an output that comprises a second predicted error rate, and wherein the second predicted error rate represents a probability that generating the data report will produce an error based on a parameter in the second category of parameters that is to be included or not included in the data report;

training, with the at least one processor, a third prediction model based on the data instances corresponding to the third category of parameters, wherein the third category of parameters comprises filters, wherein a filter is a predefined text string or a predefined numerical value, and the at least one data instance corresponding to the third category of parameters comprises a plurality of data points, wherein each data point is an integer equal to 1 or 0, wherein the integer equal to 1 represents the filter that is included in the data report and the integer equal to 0 represents the filter that is not included in the data report, wherein the third prediction model is configured to provide an output that comprises a third predicted error rate, and wherein the third predicted error rate represents a probability that generating a data report will produce an error based on a parameter in the third category of parameters that is to be included or not included in the data report;

training, with the at least one processor, an ensemble machine learning model based on an output of the first prediction model, an output of the second prediction model, and an output of the third prediction model to provide a trained ensemble machine learning model; and

generating, with the at least one processor, a runtime output of the trained ensemble machine learning model based on a runtime input, wherein the runtime output comprises a prediction of a total error rate and an indication of a predicted cause of the total error rate based on one or more parameters selected by a user, and wherein generating the runtime output of the trained ensemble machine learning model comprises:

generating a first runtime error rate with the first prediction model;

generating a second runtime error rate with the second prediction model;

generating a third runtime error rate with the third prediction model;

generating a total error rate based on the first runtime error rate, the second runtime error rate, and the third runtime error rate; and

generating an indication of a predicted cause of the total error rate with the trained ensemble machine learning model based on the one or more parameters selected by the user for a data report associated with the runtime input.

9 . The method of claim 8 , further comprising:

performing an action based on the runtime output of the trained ensemble machine learning model.

10 . The method of claim 8 , wherein receiving the dataset of the plurality of data instances comprises:

performing a normalization operation on the plurality of data instances to provide a plurality of normalized data instances, wherein each data instance comprises a plurality of data points, and wherein each data point has a value from 0 to 1.

11 . The method of claim 8 , wherein generating a runtime output of the trained ensemble machine learning model comprises:

receiving a first data instance corresponding to the first category of parameters, wherein the first data instance comprises a plurality of data points, and wherein each data point is an integer representing a first parameter that is selected by the user to be included or not included in the data report;

receiving a second data instance corresponding to the second category of parameters, wherein the second data instance comprises a plurality of data points, and wherein each data point is an integer representing a second parameter that is selected by the user to be included or not included in the data report;

receiving a third data instance corresponding to the third category of parameters, wherein the third data instance comprises a plurality of data points, and wherein each data point is an integer representing a third parameter that is selected by the user to be included or not included in the data report;

generating a first runtime error rate with the first prediction model based on the first data instance;

generating a second runtime error rate with the second prediction model based on the second data instance;

generating a third runtime error rate with the third prediction model based on the third data instance; and

generating the total error rate and the indication of the predicted cause of the total error rate with the trained ensemble machine learning model based on the first runtime error rate, the second runtime error rate, and the third runtime error rate.

12 . The method of claim 8 , wherein the first category of parameters is attributes, wherein an attribute is a text string and the at least one data instance corresponding to the first category of parameters comprises a plurality of data points, and wherein each data point is an integer representing the attribute that is to be included or not included in the data report.

13 . The method of claim 8 , wherein the second category of parameters is metrics, wherein a metric is a numerical value and the at least one data instance corresponding to the second category of parameters comprises a plurality of data points, and wherein each data point is an integer representing the metric that is to be included or not included in the data report.

14 . The method of claim 8 , wherein the third category of parameters is filters, wherein a filter is a predefined text string or a predefined numerical value, and the at least one data instance corresponding to the third category of parameters comprises a plurality of data points, wherein each data point is an integer equal to 1 or 0, and wherein the integer equal to 1 represents the filter that is included in the data report and the integer equal to 0 represents the filter that is not included in the data report.

15 . A computer program product, the computer program product comprising at least one non-transitory computer-readable medium including one or more instructions that, when executed by at least one processor, cause the at least one processor to:

receive a dataset of a plurality of data instances, wherein the plurality of data instances comprises at least one data instance corresponding to a first category of parameters, at least one data instance corresponding to a second category of parameters, and at least one data instance corresponding to a third category of parameters;

train a first prediction model based on the data instances corresponding to the first category of parameters, wherein the first category of parameters comprises attributes, wherein an attribute is a text string and the at least one data instance corresponding to the first category of parameters comprises a plurality of data points, wherein each data point is an integer representing the attribute that is to be included or not included in a data report, wherein the first prediction model is configured to provide an output that comprises a first predicted error rate, and wherein the first predicted error rate represents a probability that generating the data report will produce an error based on a parameter in the first category of parameters that is to be included or not included in the data report;

train a second prediction model based on the data instances corresponding to the second category of parameters, wherein the second category of parameters comprises metrics, wherein a metric is a numerical value and the at least one data instance corresponding to the second category of parameters comprises a plurality of data points, wherein each data point is an integer representing the metric that is to be included or not included in the data report, wherein the second prediction model is configured to provide an output that comprises a second predicted error rate, and wherein the second predicted error rate represents a probability that generating the data report will produce an error based on a parameter in the second category of parameters that is to be included or not included in the data report;

train a third prediction model based on the data instances corresponding to the third category of parameters, wherein the third category of parameters comprises filters, wherein a filter is a predefined text string or a predefined numerical value, and the at least one data instance corresponding to the third category of parameters comprises a plurality of data points, wherein each data point is an integer equal to 1 or 0, wherein the integer equal to 1 represents the filter that is included in the data report and the integer equal to 0 represents the filter that is not included in the data report, wherein the third prediction model is configured to provide an output that comprises a third predicted error rate, and wherein the third predicted error rate represents a probability that generating a data report will produce an error based on a parameter in the third category of parameters that is to be included or not included in the data report;

train an ensemble machine learning model based on an output of the first prediction model, an output of the second prediction model, and an output of the third prediction model to provide a trained ensemble machine learning model; and

generate a runtime output of the trained ensemble machine learning model based on a runtime input, wherein the runtime output comprises a prediction of a total error rate and an indication of a predicted cause of the total error rate based on one or more parameters selected by a user, and wherein, the one or more instructions that cause the at least one processor to generate the runtime output of the trained ensemble machine learning model, cause the at least one processor to:

generate a first runtime error rate with the first prediction model;

generate a second runtime error rate with the second prediction model;

generate a third runtime error rate with the third prediction model;

generate a total error rate based on the first runtime error rate, the second runtime error rate, and the third runtime error rate; and

generate an indication of a predicted cause of the total error rate with the trained ensemble machine learning model based on the one or more parameters selected by the user for a data report associated with the runtime input.

16 . The computer program product of claim 15 , wherein the one or more instructions further cause the at least one processor to:

perform an action based on the runtime output of the trained ensemble machine learning model.

17 . The computer program product of claim 15 , wherein, the one or more instructions that cause the at least one processor to generate a runtime output of the trained ensemble machine learning model, cause the at least one processor to:

receive a first data instance corresponding to the first category of parameters, wherein the first data instance comprises a plurality of data points, and wherein each data point is an integer representing a first parameter that is selected by the user to be included or not included in the data report;

receive a second data instance corresponding to the second category of parameters, wherein the second data instance comprises a plurality of data points, and wherein each data point is an integer representing a second parameter that is selected by the user to be included or not included in the data report;

receive a third data instance corresponding to the third category of parameters, wherein the third data instance comprises a plurality of data points, and wherein each data point is an integer representing a third parameter that is selected by the user to be included or not included in the data report;

generate a first runtime error rate with the first prediction model based on the first data instance;

generate a second runtime error rate with the second prediction model based on the second data instance;

generate a third runtime error rate with the third prediction model based on the third data instance; and

generate the total error rate and the indication of the predicted cause of the total error rate with the trained ensemble machine learning model based on the first runtime error rate, the second runtime error rate, and the third runtime error rate.

18 . The computer program product of claim 15 , wherein the first category of parameters is attributes, wherein an attribute is a text string and the at least one data instance corresponding to the first category of parameters comprises a plurality of data points, and wherein each data point is an integer representing the attribute that is to be included or not included in the data report.

19 . The computer program product of claim 15 , wherein the second category of parameters is metrics, wherein a metric is a numerical value and the at least one data instance corresponding to the second category of parameters comprises a plurality of data points, and wherein each data point is an integer representing the metric that is to be included or not included in the data report.

20 . The computer program product of claim 15 , wherein the third category of parameters is filters, wherein a filter is a predefined text string or a predefined numerical value, and the at least one data instance corresponding to the third category of parameters comprises a plurality of data points, wherein each data point is an integer equal to 1 or 0, and wherein the integer equal to 1 represents the filter that is included in the data report and the integer equal to 0 represents the filter that is not included in the data report.