still is not clear
is it better than chi^2 or mutual info?

i think that it's a great idea, but in my opinion you don't consider the number of cases of the different categories of example if you have two cases of 'married' and the two persons have positive class, the WOE of the married category will be very high, and maybe it's category don't have enough cases

ISSUE 01:

Apply the Mono_Bin Binning Function for Continuous Variables: Y = Target, X = Variable, n = max_bin

FIRST Variable: x1

Y = SF_Train5['y']
X = SF_Train5['x1']
max_bin = 10
force_bin = 3

WOE_01 , IV_01 = Mono_Bin(Y , X , n = max_bin)

WOE_01
The program works and produces the Output:

SECOND Variable: x4

Y = SF_Train5['y']
X = SF_Train5['x4']
max_bin = 10
force_bin = 3

WOE_04 , IV_04 = Mono_Bin(Y , X , n = max_bin)

WOE_04
The program does NOT work and produces the ValueError:

ValueError: percentiles should all be in the interval [0, 1].
if len(d2) == 1:
35 n = force_bin
---> 37 bins = df1.quantile(notmiss.X, np.linspace(0, 1, n))

I have compared the range of values within the two variables which had previously been transformed
using the Min_Max Scaler to be between 0 and 100 as well as the separation of their percentile values.

No Apparent Reason why the code should work with all my data variables except this one??

ISSUE 02:

The output table from your example produces the actual name of the column ( VAR_NAME) in the input dataframe
for which the WOE is being computed. However when I run the code on my data my table ( VAR_NAME) does not
have the actual name in the table but simply VAR for each and every variable on each run through my dataframe columns...

I would appreciate some help with fixing the above two problems.

`

test_zs.zip

`

Hi Sundar,

I am new to Python and found this code to be extremely helpful. Could you explain what this section of the code is doing in the data_vars function?

stack = traceback.extract_stack()
filename, lineno, function_name, code = stack[-2]
vars_name = re.compile(r'\((.*?)\).*$').search(code).groups()[0]
final = (re.findall(r"[\w']+", vars_name))[-1]

The reason why I ask is because I got an error in the vars_name line and I'm not sure what is happening here:

TypeError: expected string or buffer

From some quick StackOverflow searches, I believe this means it is expecting a string from "code", but it is being fed some other data type, perhaps a list. I quickly added a line to print the type of code and then the python script started working flawlessly.

stack = traceback.extract_stack()
filename, lineno, function_name, code = stack[-2]
print(type(code))
vars_name = re.compile(r'\((.*?)\).*$').search(code).groups()[0]
final = (re.findall(r"[\w']+", vars_name))[-1]

I'm a bit confused as to what happened, and why adding print(type(code)) would fix the issue. Could you provide some insight?

Thanks!
Brian