It is versatile to adapt to the linear separability of data. For a separable dataset, SVM with linear kernel performs comparably to logistic regression. Beyond this, SVM also works well for a non-separable one, if equipped with a non-linear kernel, such as RBF. For a high-dimensional dataset, the performance of logistic regression is usually compromised, while SVM still performs well. A good example could be news classification where the feature dimension is tens of thousands. In general, very high accuracy can be achieved by SVM with the right kernel and parameters. However, this might be at the expense of intense computation and high memory consumption.