In Python small floats tending to zero

python floating-point numerical-stability
Pravel picture Pravel · Sep 13, 2010 · Viewed 16.8k times · Source

I have a Bayesian Classifier programmed in Python, the problem is that when I multiply the features probabilities I get VERY small float values like 2.5e-320 or something like that, and suddenly it turns into 0.0. The 0.0 is obviously of no use to me since I must find the "best" class based on which class returns the MAX value (greater value).

What would be the best way to deal with this? I thought about finding the exponential portion of the number (-320) and, if it goes too low, multiplying the value by 1e20 or some value like that. But maybe there is a better way?

Answer

Muhammad Alkarouri picture Muhammad Alkarouri · Sep 13, 2010

What you describe is a standard problem with the naive Bayes classifier. You can search for underflow with that to find the answer. or see here.

The short answer is it is standard to express all that in terms of logarithms. So rather than multiplying probabilities, you sum their logarithms.

You might want to look at other algorithms as well for classification.

Related questions

How do I parse a string to a float or int?

In Python, how can I parse a numeric string like "545.2222" to its corresponding float value, 545.2222? Or parse the string "31" to an integer, 31? I just want to know how to parse a float str to a float, and (separately) an int …

Read More
Limiting floats to two decimal points

I want a to be rounded to 13.95. >>> a 13.949999999999999 >>> round(a, 2) 13.949999999999999 The round function does not work the way I expected.

Read More
How do I check if a string is a number (float)?

What is the best possible way to check if a string can be represented as a number in Python? The function I currently have right now is: def is_number(s): try: float(s) return True except ValueError: return False …

Read More