Welcome to BEAR Predction!

The BEAR Predction is an online tool built to perform bootstrap feature selection and ensemble classification.  

Methods:

The flowchart of BEAR Prediction is presented below. For more detailed description, please refer to our paper in BMC Biology Director (coming soon).

 

 

Input:

The BEAR Prediction accepts CSV (Comma Separated Values) files as inputs, including training data set and test (validation) data set. The input could also include a file containing true labels for test data set. A training data set is a matrix that records feature (column) values of samples (row). The last coumn in the data matrix is the class label for each sample. A test data set has the same format as the training data set, but the true sample class label is not present. Instead, the sample's true class label is provided in another txt file. An example of all three input files can be download from here.

 

Notice: most of errors during execution of the web server are caused by submitting data with wrong format. If you are running into execution errors. Please first check the format of your input data. Note that the training and test data should be in CSV (Comma Separated Values) format.

 

Output:

The output of BEAR Predction is a zipped folder consisting of the final feature selection and class prediction results and some important intermediate files and figures, including Venn diagram analysis, weigthed emsemble data sets etc.

 

*If your data is large (over 10MB), the processing time might be very long on the server. You could submit a request to us by sending us an email at BioComs.org.

 

Now, let's get started!

 

First, please provide training data set and test data set. There are two options to provide training data. Please choose either oversampling (Option 1) or no oversampling (Option 2) for training data (but not both). The personal contact information is optional, and can leave as blank.

 

Training Data (Option 1): please upload your training data here (oversampling):

Training Data (Option 2): please upload your training data here (no oversampling):

Test Data: please upload your test data here:



Name:    

Email:    

Message: