The comparison of three gradient-based approaches. Here, we briefly compare the differences, pros, and cons of a Stochastic Gradient Descent (SGD), a Batch Gradient Descent (BGD), and a Mini-Batch Gradient Descent (MBGD) methods. The description of each column is as follows,

"Algorithm" indicates the corresponding algorithm's name or approach.

"Differences" shows the difference between the corresponding method and the other methods.

"Pros" describes the advantages of the corresponding method compared to others.

"Cons" explains the shortcomings of the corresponding method compared to others.

Note the table can be sorted by clicking on the title of each column.
Algorithm	Differences	Pros	Cons	Reference
Algorithm	Differences	Pros	Cons	Reference
SGD-based SVMs	Uses one example in each iteration	Trains only one example	Updates parameters one by one Cannot be applied for large-scale problems	Lin et al. [2011]
BGD-based SVMs	Uses the whole examples in each iteration	Cumulates the update of the parameters in a batch	Unsuitable for large-scale problems	Chu et al. [2007]
BMBGD-based SVMs	Uses a part of examples in each iteration Parallelized on Spark	Suitable for large-scale problems Fewer calculation and higher accuracy Solves iterative algorithms Faster than Hadoop Constant space and time complexity per update Higher accuracy compared to SGD		Tao et al. [2014]