Automatic machine learning model generation

Abstract

A system may automatically generate a predictive machine learning model by automatically performing various processes based on an analysis of the data as well as metadata associated with the data. The system may accept a selection of data and a prediction field from the data. The system may automatically generate a set of features based on the data and may automatically remove certain features that cause inaccuracies in the model. The system may balance the data based on a representation rate of certain outcomes. The system may train and select a model based on several candidate models. The system may then perform the predictions based on the selected model and send an indication of the predictions to a user.

What is claimed is: 
     
         1 . A method for data processing, comprising:
 receiving a selection of a dataset, a prediction field from the dataset, and a set of parameters associated with the prediction field;   processing the dataset based at least in part on sampling the dataset according to a sampling rate, removing a portion of the dataset, filtering out a set of outliers from the dataset, using metadata associated with the dataset to generate features for the dataset, and segmenting the dataset into a first subset of training data and a second subset of evaluation data;   training a plurality of candidate predictive machine learning models using the first subset of training data from the processed dataset and the set of parameters associated with the prediction field;   evaluating a predictive accuracy of the plurality of candidate predictive machine learning models using the second subset of evaluation data from the processed dataset;   transmitting an indication of the plurality of candidate predictive machine learning models and the predictive accuracy of the plurality of candidate predictive machine learning models;   receiving a selection of a first predictive machine learning model from the plurality of candidate predictive machine learning models; and   predicting one or more values for the prediction field using the first predictive machine learning model.   
     
     
         2 . The method of  claim 1 , further comprising:
 determining respective scores for the plurality of candidate predictive machine learning models based at least in part on performing a statistical analysis of the predictive accuracy of the plurality of candidate predictive machine learning models.   
     
     
         3 . The method of  claim 2 , wherein the respective scores for the plurality of candidate predictive machine learning models are displayed on the user interface of the client device. 
     
     
         4 . The method of  claim 1 , further comprising:
 determining a correlation between the one or more values generated by the first predictive machine learning model and actual values from the dataset.   
     
     
         5 . The method of  claim 4 , further comprising:
 displaying the correlation.   
     
     
         6 . The method of  claim 1 , wherein the metadata indicates respective data types and characteristics of fields in the dataset. 
     
     
         7 . The method of  claim 1 , wherein the plurality of candidate predictive machine learning models are trained using parameters defined by a user of a client device. 
     
     
         8 . An apparatus for data processing, comprising:
 a processor;   memory coupled with the processor; and   instructions stored in the memory and executable by the processor to cause the apparatus to:
 receive a selection of a dataset, a prediction field from the dataset, and a set of parameters associated with the prediction field; 
 process the dataset based at least in part on sampling the dataset according to a sampling rate, removing a portion of the dataset, filtering out a set of outliers from the dataset, using metadata associated with the dataset to generate features for the dataset, and segmenting the dataset into a first subset of training data and a second subset of evaluation data; 
 train a plurality of candidate predictive machine learning models using the first subset of training data from the processed dataset and the set of parameters associated with the prediction field; 
 evaluate a predictive accuracy of the plurality of candidate predictive machine learning models using the second subset of evaluation data from the processed dataset; 
 transmit an indication of the plurality of candidate predictive machine learning models and the predictive accuracy of the plurality of candidate predictive machine learning models; 
 receive a selection of a first predictive machine learning model from the plurality of candidate predictive machine learning models; and 
 predict one or more values for the prediction field using the first predictive machine learning model. 
   
     
     
         9 . The apparatus of  claim 8 , wherein the instructions are further executable by the processor to cause the apparatus to:
 determine respective scores for the plurality of candidate predictive machine learning models based at least in part on performing a statistical analysis of the predictive accuracy of the plurality of candidate predictive machine learning models.   
     
     
         10 . The apparatus of  claim 9 , wherein the respective scores for the plurality of candidate predictive machine learning models are displayed on the user interface of the client device. 
     
     
         11 . The apparatus of  claim 8 , wherein the instructions are further executable by the processor to cause the apparatus to:
 determine a correlation between the one or more values generated by the first predictive machine learning model and actual values from the dataset.   
     
     
         12 . The apparatus of  claim 11 , wherein the instructions are further executable by the processor to cause the apparatus to:
 display the correlation.   
     
     
         13 . The apparatus of  claim 8 , wherein the metadata indicates respective data types and characteristics of fields in the dataset. 
     
     
         14 . The apparatus of  claim 8 , wherein the plurality of candidate predictive machine learning models are trained using parameters defined by a user of a client device. 
     
     
         15 . A non-transitory computer-readable medium storing code for data processing, the code comprising instructions executable by a processor to:
 receive a selection of a dataset, a prediction field from the dataset, and a set of parameters associated with the prediction field;   process the dataset based at least in part on sampling the dataset according to a sampling rate, removing a portion of the dataset, filtering out a set of outliers from the dataset, using metadata associated with the dataset to generate features for the dataset, and segmenting the dataset into a first subset of training data and a second subset of evaluation data;   train a plurality of candidate predictive machine learning models using the first subset of training data from the processed dataset and the set of parameters associated with the prediction field;   evaluate a predictive accuracy of the plurality of candidate predictive machine learning models using the second subset of evaluation data from the processed dataset;   transmit an indication of the plurality of candidate predictive machine learning models and the predictive accuracy of the plurality of candidate predictive machine learning models;   receive a selection of a first predictive machine learning model from the plurality of candidate predictive machine learning models; and   predict one or more values for the prediction field using the first predictive machine learning model.   
     
     
         16 . The non-transitory computer-readable medium of  claim 15 , wherein the instructions are further executable by the processor to:
 determine respective scores for the plurality of candidate predictive machine learning models based at least in part on performing a statistical analysis of the predictive accuracy of the plurality of candidate predictive machine learning models.   
     
     
         17 . The non-transitory computer-readable medium of  claim 16 , wherein the respective scores for the plurality of candidate predictive machine learning models are displayed on the user interface of the client device. 
     
     
         18 . The non-transitory computer-readable medium of  claim 15 , wherein the instructions are further executable by the processor to:
 determine a correlation between the one or more values generated by the first predictive machine learning model and actual values from the dataset.   
     
     
         19 . The non-transitory computer-readable medium of  claim 18 , wherein the instructions are further executable by the processor to:
 display the correlation.   
     
     
         20 . The non-transitory computer-readable medium of  claim 15 , wherein the metadata indicates respective data types and characteristics of fields in the dataset.