×
BigML is working hard to support a wide range of browsers. Your experience will be better with:
Order Instances in the Dashboard
Nov2018
Image of Order Instances in the Dashboard

The ordering instances option in BigML allows you to sort the rows of a dataset by one or more selected fields in ascending or descending order. The instances will be sorted first by the first selected field, then by the second field, and so on. You can select up to 8 different sorting fields.

This option is very useful for time series, when you have a dataset containing a date field and you need to sort your instances chronologically.

dashboard data transformations order dataset
Feature Engineering with Flatline Editor
Oct2018
Image of Feature Engineering with Flatline Editor

The BigML Flatline Editor has been upgraded to easily help you create new fields and validate existing Flatline expressions in your Dashboard. Flatline is BigML's domain-specific language for data generation and filtering, which helps you to perform an infinite number of calculations on top of your dataset fields.

BigML included a table-like dataset preview where you can easily see a sample of your instances. When you write a formula and you want to view its result, the preview only shows the fields involved in the formula. That way you can quickly check if your formula is being calculated correctly. Moreover, BigML also included a formula autocompletion so it's convenient to see which operators and dataset fields you can use while writing in the editor.

Find more information in the Datasets with the BigML Dashboard document.

dashboard data feature engineering flatline dataset
Sliding Windows in BigML
Oct2018
Image of Sliding Windows in BigML

Creating new features using sliding windows is one of the most common feature engineering techniques in Machine Learning. It is usually applied to frame time series data using previous data points as new input fields to predict the next time data points. For example, imagine we have one year of sales data to predict sales, we can use our sales field to create an infinite number of fields containing past data: last day sales, the average of last week sales, the difference between last month and this month sales, etc. To set up an sliding window in BigML you just need to choose the operation you want to apply to the instances in the window and define a window start and end.

Find more information in the Datasets with the BigML Dashboard document.

dataset dashboard Flatline feature engineering
Merge Datasets in the Dashboard
Oct2018
Image of Merge Datasets in the Dashboard

The merging datasets option in BigML allows you to include the instances of several datasets in one dataset. This functionality can be very useful when you use multiple sources of data. For example, imagine that you collect data on an hourly basis and want to create a dataset aggregating data collected over the whole day. You only need to send the new data generated each hour to BigML, create a source and a dataset for each one, and then merge all the individual datasets into one at the end of the day.

Find more information in the Datasets with the BigML Dashboard document.

dashboard data transformations merge dataset
Join Datasets in the Dashboard
Oct2018
Image of Join Datasets in the Dashboard

BigML allows you to join several datasets to combine their fields and instances based on one or more related fields between them. This is very useful when your data is scattered in two or more datasets. For example, imagine you have employee data in one dataset and department data in another dataset. You can add the department information per employee if you have a common field to join them such as department_id.

Find more information in the Datasets with the BigML Dashboard document.

dashboard data transformations SQL join dataset
Remove Duplicates in the Dashboard
Oct2018
Image of Remove Duplicates in the Dashboard

Duplicated instances in a dataset can be problematic for training Machine Learning models. For example, if you make a random split of your dataset and you take one subset for training and other for testing, it's likely that these duplicated instances appear in both subsets, which will give you an unrealistically good performance of your model. With BigML you can now easily remove the duplicated instances in your datasets with one click.

Find more information in the Datasets with the BigML Dashboard document.

dashboard data transformations remove duplicates SQL dataset
Aggregate Instances in the Dashboard
Oct2018
Image of Aggregate Instances in the Dashboard

The aggregating instances option in BigML allows you to group the rows of a dataset by a given field. This is a very common case in Machine Learning. For example, imagine you have a dataset in which each instance is a purchase but you want to make an analysis based on customers, not purchases. In this case, you need to group your instances per customer to have a customer per row instead of a purchase.

In BigML, you can also perform multiple aggregation operations on top of the dataset fields such as sum, average, maximum, minimum, count, and count distinct, among others.

Find more information in the Datasets with the BigML Dashboard document.

dashboard data transformations SQL group aggregate dataset
SQL in the BigML API
Oct2018
Image of SQL in the BigML API

The BigML team is proud to announce SQL capabilities in the BigML API . This new feature opens up an infinite number of transformations to prepare your data for Machine Learning. The ability to freely write SQL statements will be an API-only feature for now; however, we are bringing some common transformations to the Dashboard for users that prefer to transform their data in a few clicks: aggregate instances, remove duplicates, join and merge datasets. The idea is to add more options in the Dashboard on an ongoing basis; for example, the ability to order instances.

dataset SQL transformations aggregate join remove duplicates merge dashboard API
Make Scripts Public in your Organization
Sep2018
Image of Make Scripts Public in your Organization

Until now, the scripts in an Organization belonged to a single project, which meant popular scripts had to be manually added to multiple projects. Now, you can make a script public for all the projects in your Organization, so you can share common scripts like cross-validation or feature selection with all users across all projects in a single click. Moreover, you can add those public scripts to your Dashboard menus and they will automatically be visible in all the projects in a given Organization.

The script owner and Organization admins can always revert this setting by using the switch within the script info panel, and the script will only appear in the original project where it was created.

organizations collaboration projects scripts Gallery github script
Show New Fields Information
Sep2018
Image of Show New Fields Information

You can create new fields for your dataset if you choose the option "Add fields" from the dataset configuration menu in the Dashboard. Then you can either use our pre-defined options or the Flatline editor to create the field. Both options use BigML’s domain-specific language Flatline to generate the new fields.

Now, each time you create a new field for your dataset, you will be able to see the Flatline formula used to create it. You can easily copy and paste this formula to recreate a similar field for another dataset.

dataset feature engineering new fields Flatline formula
Fusion Predictions
Jul2018
Image of Fusion Predictions

One of the main goals of building a Fusion is to make predictions with it. Fusions can be used to predict categorical or numeric values. For classification problems, fusions average the per-class probabilities across all the component models. The class with the highest probability is predicted. For regression models, the final prediction is the result of averaging the per-model predicted values.

You can perform single predictions with fusions, if you want to predict just one instance; or batch predictions if you want to predict multiple instances at the same time. For single predictions you can ask for the prediction explanation, i.e., the per-field importances in the prediction. Learn more about fusions here!

predictions classification regression probabilities prediction explanation fusion
Fusion Evaluations
Jul2018
Image of Fusion Evaluations

Evaluate the performance of your Fusions to get an estimate of how good your model is at making predictions for new data. As with other supervised learning models, easily interpret the results with BigML evaluation visualizations such as the confusion matrix, the Precision-Recall, ROC, Gain and Lift curves for classification models. Since Fusion evaluations use the same performance metrics as other classification or regression models, you can quickly compare them with the BigML evaluation comparison tool. Learn more about fusions here!

evaluations classification regression supervised learning fusion
Fusions
Jul2018
Image of Fusions

BigML is releasing Fusions, the new way to combine supervised models for improved performance. BigML Fusions combine multiple decision trees, ensembles, logistic regressions, and/or deepnets and aggregate their predictions to balance out the individual weaknesses of the single models for better performance and stability. Fusions are based on the assumption that the combination of multiple models often outperforms the component single models.

BigML provides a unique and intuitive visualization for Fusions: a Partial Dependence Plot for you to further inspect the Fusion results. Learn more about fusions here!

supervised classification regression dashboard partial dependence plot models logistic regression deepnets ensembles fusion
New Text Analysis Options
Jul2018
Image of New Text Analysis Options

BigML is happy to announce new options to analyze your text fields. With the newly added capabilities, supervised and unsupervised models will be able to identify more patterns in your text data. The configuration options include:

  • 15 new languages! BigML can now process text in 22 different languages: Arabic, Catalan, Chinese, Czech, Danish, Dutch, English, Farsi/Persian, Finish, French, German, Hungarian, Italian, Japanese, Korean, Polish, Portuguese, Turkish, Romanian, Russian, Spanish, and Swedish.
  • The maximum n-gram size to consider for your text analysis: bigrams, trigrams, four-grams and five-grams.
  • New stop words removal techniques like the ability to remove stop words for all languages and the degree of aggressiveness for stopword removal.
  • Stemming for the new languages.
  • Filters to exclude certain groups of uninteresting words from your models such as HTML keywords, non-dictionary words, numeric digits or non-language characters. You can also choose to exclude unigrams from your text and keep only n-grams that include more than one word.

You can configure these options from your source so they will be taken into account by all your models. Moreover, as you iterate your model, you can easily configure these options for your topic models regardless of your original source configuration.

models supervised unsupervised text fields topic models sources miscellaneous
The BigML Zapier App
Jun2018
Image of The BigML Zapier App

The BigML Zapier app allows you to easily automate your Machine Learning workflows without any coding. Import your data in real-time from the most popular web apps and the BigML app will automatically make predictions for you as your new data is being generated. Then you can simply choose which service or app you want to send the predictions to, and the BigML Zapier app will take care of integrating those predictions into your processes. See some workflow examples and try it for free today!

workflows automation supervised models unsupervised models predictions miscellaneous
OptiML: Automatically Get the Optimal Model
May2018
Image of OptiML: Automatically Get the Optimal Model

OptiML is an optimization process for model selection and parametrization that automatically finds the best supervised model to help you solve classification and regression problems.

Using Bayesian Parameter Optimization, OptiML creates and evaluates hundreds of supervised models (decision trees, ensembles, logistic regressions, and deepnets) and returns a list of the best models for your data. Eliminating the need for manual, trial-and-error based exploration of algorithms and parameters, OptiML saves significant time and provides improved performance for Machine Learning practitioners of all levels.

optiml classification regression models ensembles logistic regression deepnets optimization evaluations cross-validation parametrization model selection miscellaneous
Prediction Explanation
Apr2018
Image of Prediction Explanation

For any classification or regression model, you can now request an explanation for your predictions. The explanation represents the most important factors considered by the model in the prediction given the input values. Therefore, each input value will have an associated importance where the importances across all input values sum up to 100%.

While the prediction explanation for decision tree models is calculated using the prediction path, for other models (ensembles, logistic regression, and deepnets) the prediction explanation is calculated by aggregating the results of over a thousand distinct predictions that use random perturbations of the input data.

To read more about the prediction explanation with the BigML Dashboard please read the sections "Prediction Explanation" of the Classification and Regression document. To get the prediction explanation with the BigML API you need to use the "explain":true argument as detailed in API documentation.

supervised classification regression single predictions field importances prediction
BigML in Chinese
Feb2018
Image of BigML in Chinese

The BigML Team is excited to release the BigML Dashboard in Chinese. Since its very beginning, BigML has strived to make Machine Learning Beautifully Simple for Everyone (机器学习美观简单人人用). Today our journey reached another milestone by allowing over 1 billion people to use the BigML platform in their native language.

You can watch this video to check out the BigML Dashboard in Chinese

chinese languages dashboard web
Organizations
Jan2018
Image of Organizations

Machine Learning projects usually involve several members from different teams that need to work on the same project and/or visualize the same resources. Now, BigML offers organizations, a convenient collaborative workspace that makes your life easier and your work more efficient, where several users can work on the same projects from different accounts at different permission levels. Break down silos and let us help you to adopt Machine Learning across your entire corporate structure.

Create a new organization or learn more about it.

organizations collaboration projects permissions roles web
New voting strategies for ensembles
Jan2018
Image of New voting strategies for ensembles

As you may know, ensembles are composed of several decision trees. For non-boosted ensembles, each single tree returns a prediction given the input data. These predictions need to be combined so the ensemble returns a single prediction for each new instance.

For the previous years, BigML has been offering what we called "combiners": plurality, confidence weighted and probability weighted. Each of them provided a different strategy to combine single tree predictions. Now BigML is improving these strategies for ensembles and homogenizing the predictions across all resources, offering three new options that will replace our old combiners forever:

  • Probabilities: Averages the per-class probability distributions for all trees in the ensemble and predicts the class with the highest probability. For regression ensembles, the global prediction is the mean of the individual predictions.
  • Confidences: Averages the per-class confidence distributions for all trees in the ensemble and predicts the class with the highest confidence. For regression ensembles, the global prediction is the mean of the individual predictions weighted by the expected error.
  • Votes: Gives one vote to each model in the ensemble. For classification models, the category with the majority of votes wins. For regression models, the global prediction is the mean of the individual predictions (it gives the same results as the probability strategy).

You can find these new options among the prediction arguments in the BigML API under the name of operating_kinds. The old combiners will also still be available from the API for a while.

probabilities confidences votes evaluations batch predictions ensembles decision forests combiners prediction
Operating thresholds for Single Predictions
Jan2018
Image of Operating thresholds for Single Predictions

Set a threshold for the positive class when creating single predictions for any classification model. This technique is very useful for imbalanced datasets where one or a few classes are the majority classes. In these cases, models tend to predict the majority classes at the expense of the minority class that is usually the class of interest (called the positive class).

By setting a threshold for the positive class, this class will be predicted if its probability (confidence or votes depending on the type of measure you select) is greater than the established threshold, otherwise the following class with the highest probability (confidence or votes) will be predicted instead.

Whether you build a single decision tree, an ensemble, a logistic regression or a deepnet, you can select a positive class and set a threshold for a given prediction.

predictions probabilities confidences thresholds positive class imbalanced prediction miscellaneous
Show Probabilities for Models
Jan2018
Image of Show Probabilities for Models

BigML has been always using confidences (a pessimistic approach) to measure the certainty of a given prediction. Now you can decide to also see probabilities for model predictions. The fundamental difference between both measures is that probabilities don't penalize a lower number of instances in a given node so heavily.

You can see how the three different model visualizations change by playing with the buttons to show confidences or probabilities at the top of the model view.

models visualization predictions probabilities confidences model
Normalize repeats for Anomaly Detection
Dec2017
Image of Normalize repeats for Anomaly Detection

Improve the anomalies found in your dataset using the new parameter "Normalize repeats". If you enable this parameter, Anomalies will take into account not only the different values but also the frequency of repeated (or very similar) data points.

For example, if you had many missing values in your data, it may happened that the Anomaly detector identified the instances with missing values as highly anomalous regardless whether missing values were the rule instead of the exception. By using "Normalize repeats", instances with missing data will not appear among the top anomalies in this case.

anomalies frequency score anomaly
Select date-time fields for Time Series
Dec2017
Image of Select date-time fields for Time Series

You can now select the date-time field from the dataset to plot your time series data in the Dashboard. Be aware that your instances must be chronologically sorted in the dataset to select the date-time field.

date-time fields dashboard visualization timeseries
Deepnet Predictions
Oct2017
Image of Deepnet Predictions

One of the main goals of any BigML resource is making predictions, and Deepnets are no exception. Deepnets can be used to predict categorical or numeric values. As Deepnets have more than one layer of nodes between the input and the output layers, the output will be the network’s prediction. In the case of categorical objective fields, an array of per-class probabilities will be returned, while a single, real value will be predicted for regression problems. You can perform single predictions, if you want to predict just one instance; or batch predictions if you want to predict multiple instances at the same time.

deepnets classification regression probabilities prediction
Deepnet Evaluations
Oct2017
Image of Deepnet Evaluations

Evaluate the performance of your Deepnets to get an estimate of how good your model will be at making predictions for new data and easily interpret the results with BigML evaluation visualizations. As with other supervised learning models on BigML, Deepnets include the confusion matrix and the ROC and Precision-Recall curves to help you solve classification problems. Furthermore, quickly compare the performance of your different Deepnets and other models built using different algorithms with the BigML evaluation comparison tool.

deepnets classification regression comparison tool roc curve auc confusion matrix evaluation
Deepnets
Oct2017
Image of Deepnets

BigML is proud to announce Deepnets, an optimized version of Deep Neural Networks, the machine-learned models loosely inspired by the neural circuitry of the human brain. Deepnets are state-of-the-art in many important supervised learning applications. To avoid the difficult and time-consuming work of hand-tuning the algorithm, BigML’s unique implementation of Deep Neural Networks offers first-class support for automatic network search and parameter optimization. BigML makes it easier for you by searching over all possible networks for your dataset and returning the best network found to solve your problem. Thus, non-experts can train deep learning models with results matching that of top-level data scientists.

supervised classification regression evaluation dashboard predictions partial dependence plot neural networks deep learning miscellaneous deepnet
Time Series: Model Decomposition
Sep2017
Image of Time Series: Model Decomposition

BigML recently launched Time Series, a sequentially indexed representation of your historical data commonly used for predicting stock prices, sales forecasting; website traffic, production, and inventory analysis as well as weather forecasting among other use cases.

BigML implements exponential smoothing methods which are able to learn multiple models out of the training data by using different combinations of the three essential model components: the level, the trend, and the seasonality. Now, you can decompose your models to display each of these components plotted in a separate chart. For a detailed explanation of each component, please visit the dedicated release page, where you will find a series of six blog posts about Time Series, the BigML Dashboard and API documentation, the webinar slideshow as well as the full webinar recording.

seasonality level trend dashboard decomposition models timeseries
Time Series Forecasts
Jul2017
Image of Time Series Forecasts

You can use your Time Series models to make predictions, which are called Forecasts. With Time Series Forecasts, you can easily forecast events in short or longer time horizons. You can also employ a Time Series model to forecast the future values of multiple objective fields. Along with each forecasted data point, BigML generates an error interval as a measure of the quality of your forecast. The interval indicates the lower and upper bounds within which the forecast will fall with a 95% confidence.

forecasts time series intervals horizon miscellaneous timeseries
Time Series Evaluations
Jul2017
Image of Time Series Evaluations

You can easily evaluate the performance of your Time Series models. For this, you need to use two different subsets of data: one for training and the other one for testing. BigML represents your test data and the model forecasts in a chart, so you can visually analyze the goodness-of-fit of your Time Series models.

You will also see multiple performance metrics such as the Mean Absolute Error (MAE), the Mean Squared Error (MSE), the R squared, the Symmetric Mean Absolute Percentage Error (SMAPE), the Mean Scaled Error (MSE), and the Mean Directional Accuracy (MDA). You can find an explanation for each metric in the 6th chapter of the Time Series documentation.

evaluations time series performance metrics MAE MSE R squared SMAPE MDA miscellaneous timeseries
Time Series Models
Jul2017
Image of Time Series Models

BigML is proud to launch Time Series, a sequentially indexed representation of your historical data that can be used to forecast future values of numerical properties. This is a versatile method often used for predicting stock prices, sales forecasting, website traffic, production and inventory analysis, and weather forecasting, among many other use cases.

A Time Series model needs to be trained with numeric fields containing a time-ordered sequence of regularly spaced data points in time. BigML implements exponential smoothing methods which are able to forecast time-based data with complex trends and seasonal patterns. BigML generates multiple models behind the scenes so you can select the best-performing ones. You can find a detailed explanation in the Time Series documentation.

time series forecast numeric fields supervised learning miscellaneous timeseries
Compare Multiple Evaluations
Jul2017
Image of Compare Multiple Evaluations

Any classification problem can be solved by using different supervised learning algorithms and using different configurations per algorithm as you iteratively improve your models. BigML brings to the Dashboard an easy and visual way to compare your models and decide which one performs better. You can easily select the models, the positive class, the metrics and any of the ROC curve, the precision-recall curve, the gain curve, or the lift curve for your comparison. You can also rank your models by the Area Under the Curve (AUC), K-S statistic, Kendall's Tau, or Spearman's Rho.

classification models ensembles logistic regression evaluation comparison roc curve precision-recall curve lift curve gain curve auc auch probability threshold kendall's tau spearman's rho k-s statistic evaluation
Evaluation Curves for Classification Models
Jul2017
Image of Evaluation Curves for Classification Models

Evaluating the performance of your Machine Learning models is one of the most important steps in the predictive process. BigML is releasing a new Dashboard visualization for evaluations, which includes new performance metrics to make it easier to assess the performance of your classification models. Now, you can use the popular ROC curve to understand the trade-off between sensitivity and specificity for each possible threshold as well as the precision-recall curve, the gain curve, and the lift curve. Moreover, you can find new metrics that measure the overall predictive performance of your models for the selected positive class such as the Area Under the Curve (AUC), the Area Under the Convex Hull (AUCH), the K-S statistic, the Kendall's Tau, and the Spearman's Rho coefficients.

roc curve precision-recall curve gain curve lift curve auc auch probability threshold kendall's tau spearman's rho k-s statistic evaluation
Resource Configuration Information
Jul2017
Image of Resource Configuration Information

To solve Machine Learning problems, you usually need several iterations that employ different algorithms and configurations to build your final models and workflows. Now, BigML makes it even easier and faster for you to find the right resources at a glance from among many that belong to the same project by listing the values of your configured parameters for each resource.

configuration parameters resoruces names API Dashboard miscellaneous
Select the Consequent for Associations
Mar2017
Image of Select the Consequent for Associations

Associations are powerful means in finding strong correlations among your dataset values. However, depending on your case, you may not always be interested in finding the strongest relationships but only the rules that meet certain conditions instead. Specifying your conditions to zero in on your rules of interest is now easier than ever. Simply select your data field of interest and one or more field values for the consequent part of the rule, and you will obtain the relevant associations in no time.

associations consequent antecedent fields dashboard api association
Select the Consequent for Associations
Mar2017
Image of Select the Consequent for Associations

Associations are powerful means in finding strong correlations among your dataset values. However, depending on your case, you may not always be interested in finding the strongest relationships but only the rules that meet certain conditions instead. Specifying your conditions to zero in on your rules of interest is now easier than ever. Simply select your data field of interest and one or more field values for the consequent part of the rule, and you will obtain the relevant associations in no time.

associations consequent antecedent fields dashboard api association
Boosted Trees Predictions
Mar2017
Image of Boosted Trees Predictions

The ultimate goal of creating any supervised learning model is to get a prediction for new intstances. Like other supervised models, Boosted Trees offer Single Predictions to predict a given single instance and Batch Predictions to predict multiple instances simultaneously. Instead of returning a single class along with its confidence, Boosted Trees return a set of probabilities for all the classes in the objective field which is visible in the predictions histogram.

ensemble classification regression single predictions batch predictions dashboard api boosted trees
Boosted Trees
Mar2017
Image of Boosted Trees

The BigML team is proud to announce Boosted Trees, the third ensemble-based strategy that BigML provides to help you easily solve your classification and regression problems. Together with Bagging and Random Decision Forests, Boosted Trees make for a powerful combination available both via the BigML Dashboard and our REST API. This well-known technique is an ensemble of several single models, where each tree improves the mistakes made by the previously grown tree. It is one of the best performing Machine Learning methods to solve complex real-world problems.

ensemble classification regression partial dependence plot boosting boosted trees dashboard API
Partial Dependence Plots for Models
Feb2017
Image of Partial Dependence Plots for Models

As a complement to our popular decision trees visualization and the sunburst, we are launching a third view for your models: the Partial Dependence Plot. This heatmap chart also allows you to analyze the marginal impact of each input field on predictions for classification and regression models built by using ensembles and logistic regressions.

supervised classification regression visualization dashboard predictions model miscellaneous
Batch Deletion for Resources
Feb2017
Image of Batch Deletion for Resources

Solving a Machine Learning problem is an iterative process that requires the creation of a great number of intermediary datasets, models, evaluations and predictions to get the final model. Now, BigML simplifies it keeping your account organized and up-to-date by allowing the deletion of multiple resources at the same time. Just click the deletion icon found in the resources listing in the Dashboard, and select the resources to be deleted.

deletion resources api dashboard miscellaneous
E-mail Notifications for Scripts
Feb2017
Image of E-mail Notifications for Scripts

Asynchcronous WhizzML script executions can take some time to finish to fulfill complex Machine Learning workflows they implement. You no longer need to check your execution repeatedly to see if your results are ready. This new option let's you specify that you want to be notified by e-mail once the execution finishes, while you concentrate on other tasks.

whizzml scripts executions notifications
Scriptify: Reify Complex Workflows
Jan2017
Image of Scriptify: Reify Complex Workflows

Furthering our obsession to speed up your Machine Learning processes, we have incorporated Scriptify into your 1-click menu options. Now, you can automatically regenerate any BigML resource (models, evaluations, predictions, etc) with a single click. Scriptify creates a script that contains all the workflow information end-to-end (from configuration parameters to resources created). You can precisely repeat the processing steps of any original Machine Learning resource to your heart's desire!

whizzml worklows automation resources 1-click actions
Shared Resources Cloning
Jan2017
Image of Shared Resources Cloning

Now, you can easily clone datasets, models and scripts, from other users into your BigML account. Provided that a user shares a resource using the sharing link and the cloning capability is enabled, any other user with access to the link will be able to include this resource in their BigML account.

This new feature will allow you to fully use the shared resources. For example, when another user shares a dataset using the sharing link, it is in "view only" mode, so you can not perform any actions such as creating new models, exporting it, sampling it, etc. Now, by cloning it, you will be able to perform all BigML actions available for datasets.

resources cloning sharing link dashboard web
Stats Computation for Logistic Regression
Jan2017
Image of Stats Computation for Logistic Regression

These new Dashboard statistics allow you to introspect the predictive power of your model by revealing the significance of each coefficient estimate. BigML computes the likelihood ratio to test how well the model fits your data along with the p-value, confidence interval, standard, error and Z score for each coefficient.

Learn more about the Logistic Regression statistics in the Dashboard documentation.

logistic regression dashboard supervised stats p-value z score standard error confidence intervals likelihood ratio logisticregression
Association Predictions: Association Sets
Jan2017
Image of Association Predictions: Association Sets

BigML is bringing predictions for Associations to the Dashboard. Association Sets allow you to pinpoint the items which are most strongly associated with your input data. For example, given a set of products purchased by a person, what other products are most likely to be bought?

All the predicted items will be ranked according to a similarity score, and they will be displayed in a table view. You can also visualize each predicted rule in a Venn diagram to get a sense of the correlation strength between the input data and the predicted items. Read more about Association Sets in the 8th chapter of the Associations documentation.

associations predictions itemsets association rules associationset
BigML Certifications
Jan2017
Image of BigML Certifications

We are happy to announce BigML Certifications, for organizations and professionals that want to master BigML to successfully deliver real-life Machine Learning projects. These courses are ideal for software developers, system integrators, analysts, or scientists, to boost their skill set and deliver sophisticated data-driven solutions. We offer two separate courses, each of them consisting of 4 weekly online classes of 3 hours each:

  • Certified Engineer: all you need to know about advanced modeling, advanced data transformations, and how to use the BigML API (and its wrappers) in combination with WhizzML to build and automate your Machine Learning workflows.

  • Certified Architect: learn how to implement your Machine Learning solutions so they are scalable, impactful, capable of being integrated with third-party systems, and easy to maintain and retrain.

If you successfully pass the certification exam, BigML will award you with a diploma. In addition, BigML Certified Partners will receive business referrals that help them source new Machine Learning projects.

courses modeling api supervised unsupervised whizzml data transformations engineer architect miscellaneous
Partial Dependence Plot for Ensembles
Dec2016
Image of Partial Dependence Plot for Ensembles

This new visualization for ensembles, commonly known as Partial Dependence Plot, allows you to visualize the impact that a set of fields have on predictions. You will be able to determine which fields are most relevant for ensemble predictions and how sensitive your ensemble predictions are to their different values.

The chart displays a heatmap representation of your predictions based on different values of the two selected fields in the axes regardless of the rest of the fields used to train your ensemble. You can select any categorical or numeric field for the axes and configure the values for the rest of the input fields by using the fields inspector panel on the right.

supervised classification regression visualization dashboard predictions ensemble
Batch Field Importances
Dec2016
Image of Batch Field Importances

This feature enables you to include the field importances in your batch predictions, i.e., a set of percentages indicating how much each field in your dataset contributed to the prediction of a given instance. You can include those values in your output file and dataset either with BigML Dashboard or the API. This will give you a better understanding of your predictions as it will reveal which are the most relevant fields factoring in a given prediction.

supervised predictions batch predictions regression classification models ensembles api dashboard prediction
Topic Distributions
Nov2016
Image of Topic Distributions

Topic Models assume that each document exhibits a mixture of topics. The main goal of creating a Topic Model is to discover the topic importances for a given document. For example, a document may be 70% about "Machine Learning", 20% about "stock market" and 10% about "startups".

Topic Distributions allow you to make predictions for a single data instance, and Batch Topic Distributions help predict the same for multiple instances simultaneously. Based on a given Topic Model, BigML Topic Distributions provide a set of probabilities for each data instance (one probability per topic), which indicate the relative relevance of all topics for that instance.

TopicModels TopicDistributions BatchTopicDistributions TopicProbabilities API Dashboard Fall2016 Unsupervised topicmodel
Topic Models
Nov2016
Image of Topic Models

The BigML team has brought Topic Models to the API and the Dashboard as part of Fall 2016 release. Topic Models are an optimized implementation of Latent Dirichlet Allocation, a probabilistic unsupervised learning method that determines the topics underlying a collection of documents.

Topic Models' main application areas include browsing, organizing and understanding large amounts of unstructured text data, which can be very useful for information retrieval tasks, collaborative filtering or content recommendation use cases among others.

BigML provides two original visualizations that accompany its implementation so you can better inspect your Topic Model:

  • Topic Map: get an overview of your topic importances and their thematic closeness.
  • Term Chart: get an overview of the main terms that make up your found topics.
TopicModels Topics TermChart TopicMap API Dashboard Fall2016 Unsupervised topicmodel
Flatliner code editor & evaluator
Apr2015
Image of Flatliner code editor & evaluator

Flatline is BigML’s Lisp-like language that enables you to programmatically perform an array of data transformations, including filtering and new field generation. Flatliner is a handy code editor (available in our Labs section) that helps you test your Flatline expressions.

dataset labs filter transform sample flatline miscellaneous
Evaluation Comparison
Mar2015
Image of Evaluation Comparison

You can now compare multiple evaluations against a test set in a ROC space. The graph can then be downloaded as a .PNG image, and the performance measures can be exported as a .csv for further analyses.

comparison auc evaluation labs
Google Integration
Jan2015
Image of Google Integration

With the Winter Release, you'll now be able to add sources to BigML through Google Cloud Storage and Google Drive, similar to our prior integrations with Dropbox and Azure Data Marketplace. You can also now log into BigML using your Google ID.

google datasources google drive google cloud storage winter2015 miscellaneous
Projects
Jan2015
Image of Projects

We're happy to introduce Projects to help you organize your machine learning resources. You only have to create a new project using the web interface or the API resource and update a new source to this project. All the new resources created from this source will be associated to the same project.

winter2015 projects miscellaneous
Dataset Comparison
Jan2015
Image of Dataset Comparison

This is another simple but useful application we have released into our new BigML Labs. It allows users to compare side by side two different datasets. Check it out here.

dataset winter2015 labs web
Sample Service
Dec2014
Image of Sample Service

BigML's new Sample Service provides fast access to datasets that are kept in an in-memory cache which enables a variety of sampling, filtering and correlation techniques. We have leveraged this new service to create a Dynamic Scatterplot visualization that we've released into BigML Labs.

sample dataset viz visualization winter2015 labs api
BigML Labs
Dec2014
Image of BigML Labs

Our team is constantly working on innovative applications built on top of BigML's API. We're now unveiling several of these in early access through our BigML Labs.

labs winter2015 miscellaneous
G-means Clusters
Dec2014
Image of G-means Clusters

This latest addition to BigML's unsupervised learning algorithms is ideal for when you may not know how many clusters you wish to build from your dataset.

gmeans winter2015 cluster
Cluster summary report
Oct2014
Image of Cluster summary report

Now you can download a Summary Report for your BigML Clusters. This report will inform you on the distribution of data across your clusters, as well as the associated features and data distances.

report cluster
BigML Comes to Australia and New Zealand
Sep2014
Image of BigML Comes to Australia and New Zealand

BigML is very pleased to announce that we've launched this new website to better serve our customers in Australia & New Zealand. This site will contain all of the content and functionality of our https://bigml.com site, but will provide faster performance as well as some localized content (e.g., local events and local training opportunities). Read more about in this blog post.

australia new zealand opening miscellaneous
Anomaly Detector
Sep2014
Image of Anomaly Detector

BigML makes it easy to build a top-performing anomaly detector that will help you identify instances in your dataset that do not conform to a regular pattern.

fraud detection summer2014 anomaly
Batch Anomaly Scores
Sep2014
Image of Batch Anomaly Scores

You can quickly score multiple lines of data through BigML's Batch Anomaly Score. The output can be downloaded as a .csv and/or you can use it to automatically create a new dataset.

batches fraud detection anomalyscore summer2014
Anomaly Score
Sep2014
Image of Anomaly Score

You can score individual data points against your anomaly detector by using the web interface. Simply input the variables and BigML will provide you with an anomaly percentage (a higher score reflects greater anomaly).

score fraud detection anomalyscore summer2014
New dataset from batch prediction output
Sep2014
Image of New dataset from batch prediction output

Batch predictions are a powerful way to score likely outcomes on multiple rows of data. You can now create a new dataset directly from the batch prediction output (in addition to getting the output as a .csv file).

batch prediction batch centroid batch anomalyscore workflow summer2014 dataset
Models from clusters
Aug2014
Image of Models from clusters

Now you can automatically create a model for each cluster that will not only help you better understand the cluster, but also use it to classify new instances.

model summer2014 cluster
Modeling with missing splits
Aug2014
Image of Modeling with missing splits

As we know that cleaning up data might be hard and having all the input data handy at prediction time is important, we have built a new option to create models that will generate predicates that explicitly deal with missing values.

missing values ensembles summer2014 model
Online predictions
Jul2014
Image of Online predictions

New client-side predictions make it easier than ever to explore the influence of each field in your models, ensembles or clusters. In addition, we are open sourcing the related Javascript libraries so you can leverage this functionality to build very powerful and dynamic apps and web services.

free client-side javascript summer2014 prediction
Fast ensembles
Jul2014
Image of Fast ensembles

We have refined the way the models of an ensemble are built to save a great amount of time in data transportation. This will dramatically speed up creation of your ensembles.

fast summer2014 ensemble
Sending Request...
Sending Request...