Systems and techniques for predictive data analytics

Abstract

Systems and techniques for predictive data analytics are described. In a method for selecting a predictive model for a prediction problem, the suitabilities of predictive modeling procedures for the prediction problem may be determined based on characteristics of the prediction problem and/or on attributes of the respective modeling procedures. A subset of the predictive modeling procedures may be selected based on the determined suitabilities of the selected modeling procedures for the prediction problem. A resource allocation schedule allocating computational resources for execution of the selected modeling procedures may be generated, based on the determined suitabilities of the selected modeling procedures for the prediction problem. Results of the execution of the selected modeling procedures in accordance with the resource allocation schedule may be obtained. A predictive model for the prediction problem may be selected based on those results.

What is claimed is: 
     
         1 - 30 . (canceled) 
     
     
         31 . A predictive modeling apparatus comprising:
 a memory configured to store a machine-executable module encoding a predictive modeling procedure, wherein the predictive modeling procedure comprises a plurality of tasks, wherein the machine-executable module comprises a directed graph representing dependencies between the tasks, and wherein the plurality of tasks comprises at least one pre-processing task and at least one model-fitting task; and   at least one processor configured to execute the machine-executable module, wherein executing the machine-executable module causes the apparatus to perform the predictive modeling procedure, including:   manipulating input data, comprising performing the pre-processing task on the input data; and   performing the model-fitting task, comprising:
 generating, from the pre-processed input data, training data and testing data; 
 selecting a predictive model from a plurality of predictive models based on respective suitability scores associated with each of the plurality of predictive models, the suitability score for a predictive model indicating a suitability of the predictive model for a prediction problem represented by the input data; 
 following selection of the predictive model, fitting the selected predictive model to the training data; and 
 testing the fitted model on the testing data. 
   
     
     
         32 . The predictive modeling apparatus of  claim 31 , wherein the suitability score for a predictive model is determined prior to selecting the predictive model, based on one or more attributes of the predictive model and one or more characteristics of the input data representing the prediction problem. 
     
     
         33 . The predictive modeling apparatus of  claim 31 , wherein the fitted predictive model is operable to generate predicted outcomes of instances of the prediction problem based on data representing instances of the prediction problem. 
     
     
         34 . The predictive modeling apparatus of  claim 31 , wherein the directed graph represents data flow dependencies between the tasks. 
     
     
         35 . The predictive modeling apparatus of  claim 34 , wherein the directed graph represents conditional data flow dependencies between the tasks. 
     
     
         36 . The predictive modeling apparatus of  claim 31 , wherein the directed graph represents control flow dependencies between the tasks. 
     
     
         37 . The predictive modeling apparatus of  claim 36 , wherein the directed graph represents conditional control flow dependencies between the tasks. 
     
     
         38 . The predictive modeling apparatus of  claim 31 , wherein the plurality of tasks comprises a first task, and wherein at least one node or edge of the directed graph represents a conditional data flow operation performed on data generated by the first task. 
     
     
         39 . The predictive modeling apparatus of  claim 38 , wherein performing the conditional data flow operation comprises conditionally directing at least a portion of the data generated by the first task from a first node of the graph representing the first task to a second node of the graph representing a second task based on a result of evaluating a conditional expression. 
     
     
         40 . The predictive modeling apparatus of  claim 31 , wherein the plurality of tasks comprises a first task, and wherein at least one node or edge of the directed graph represents a data filtering operation, a data transformation operation, and/or a data partitioning operation performed on data generated by the first task. 
     
     
         41 . The predictive modeling apparatus of  claim 31 , wherein the plurality of tasks comprises a first task and a second task, and wherein at least one node or edge of the directed graph represents a data combining operation performed on first data generated by the first task and second data generated by the second task. 
     
     
         42 . The predictive modeling apparatus of  claim 31 , wherein the pre-processed input data comprise at least one data set, wherein generating the training data comprises obtaining a first subset of the data set, and wherein generating the testing data comprises obtaining a second subset of the data set. 
     
     
         43 . The predictive modeling apparatus of  claim 42 , wherein performing the predictive modeling procedure further comprises performing cross-validation of the predictive model. 
     
     
         44 . The predictive modeling apparatus of  claim 31 , wherein performing the predictive modeling procedure further comprises performing nested cross-validation of the predictive model. 
     
     
         45 . The predictive modeling apparatus of  claim 31 , wherein the at least one processor is further configured to deploy the fitted model. 
     
     
         46 . The predictive modeling apparatus of  claim 45 , wherein deploying the fitted model comprises generating a plurality of predictions by applying the fitted model to other data representing instances of the prediction problem, wherein the input data do not comprise the other data. 
     
     
         47 . The predictive modeling apparatus of  claim 31 , wherein performing the pre-processing task on the input data comprises cleansing the input data, performing feature selection on the input data, and/or performing feature engineering on the input data. 
     
     
         48 . The predictive modeling apparatus of  claim 47 , wherein cleansing the input data comprises imputing one or more missing values of one or more variables in the input data, detecting and replacing one or more outlier values of one or more variables in the input data, and/or normalizing values of one or more variables in the input data. 
     
     
         49 . The predictive modeling apparatus of  claim 47 , wherein performing feature selection on the input data comprises selecting one or more variables in the input data as features for the predictive model. 
     
     
         50 . The predictive modeling apparatus of  claim 47 , wherein performing feature engineering on the input data comprises applying a transformation to one or more variables in the input data to generate a feature for the predictive model.