Constructing the term-document matrix using .csr_matrix

from import csr_matrix

docs = [["hello", "world", "hello"], ["goodbye", "cruel", "world"]]

indptr = [0] # Stores the row offsets

indices = [] # stores the column number of the element in data (column numbers can be repeated)

data = [] # Stores non-zero data elements

vocabulary = {} # key is the word vocabulary, value is the column number

for d in docs: # Iterate over each document

for term in d: # Iterate over each vocabulary term of the document

# setdefault If the term does not exist, then the new term and his column

# number len(vocabulary) to add to the dictionary, return his number;

# If term exists, leave it unfilled and return the number that already exists

index = (term, len(vocabulary))

(index)

(1)

(len(indices))

csr_matrix((data, indices, indptr), dtype=int).toarray()