I assume that you are using MATLAB 2016a version. I have a small code snippet that I think is your concern.
This would display the z1 as a fraction. Just convert the z1 variable to a floating number precision. You may use single or double point precision.
This should return the result in a numeric format with the precision according to the double or single data type.
In case you can afford to stay in the symbolic domain, you may use the variable-precision arithmetic. The result will be a symbolic data and not a numeric one.
Hope it helps.