The best way is to lower the casing to lowercase before calling the minhash function. This means casing will affect the result of the similarity matching. The function we’ve just implemented is case sensitive. I obtained the following result: 0.1 0.5384615384615384 Case sensitivity

print(minhash('welcome to medium', 'medium is a publishing platform')) print(minhash('i have a cat', 'i have a dog')) Let’s test the function out to see the similarity between two input sentences. Your final function should be as follows: try: score = jaccard_distance(shingles(input_question), shingles(compare_question)) except ZeroDivisionError: print('ZeroDivisionError') It’s advisable to wrap it inside a try except call. Once you’re done, we’ll begin to implement the main code. jaccard_distance = lambda seta, setb: len(seta & setb)/float(len(seta | setb)) Create an anonymous function that accepts two inputs. The next step is to calculate the Jaccard distance based on the formula given above. The order might be different, but the elements should be the same if you used the same input. You can test out the results by running the following code: shingles = lambda s: set(s for i in range(len(s)-2)) print(shingles('Hello, there ! How are you ?'), shingles('早上好，在干嘛呢？')) You can resolve it by implementing your own way to split the text into elements or pad extra spaces to the input text if the length is less than 3. It has a big issue when it comes to languages such as Chinese and Japanese due to the fact a word can be made up of just one character. Please note this anonymous function is based on the original implementation under the jProcessing module. shingles = lambda s: set(s for i in range(len(s)-2)) Then, we write the following anonymous function to convert the input text into elements of three characters. Initialize a score variable, and set it to 0. def minhash(input_question, compare_question): Feel free to name them anything as you like. I’ll just name them input_question and compare_question. This function accepts two input string parameters. Create a new function called minhash in your Python file.

Jaccard similarity python