I have the matlab code below for the attached question. I need help running this in python. can I get some help translating these to python codes/algorithm.

% Poisson penalty W = rand(V,K); H = rand(K,N); obj = []; for ite = 1:150 Wn = W./repmat(sum(W,1),V,1); X2 = X./(W*H); H = H.*(Wn*X2) + eps; Hn = H./repmat(sum(H,2),1,N); X2 = X./(eps+W*H); W = W.*(X2*Hn) + eps; obj(end+1) = sum(sum(X2)) - sum(sum(X.*log(X2+eps)))

end.

Question: end

In this problem you will factorize an N x M matrix X into a rank-K approximation WH, where W is N x K, H is K x M and all values in the matrices are nonnegative. Each value in W and H can be initialized randomly to a positive number, e.g., from a Uniform(1,2) distribution. The data to be used for this problem consists of 8447 documents from The New York Times. (See below for how to process the data.) The vocabulary size is 3012 words. You will need to use this data to construct the matrix X, where Xij is the number of times word i appears in document j. Therefore, X is 3012x8447 and most values in X will equal zero. a) Implement and run the NMF algorithm on this data using the divergence penalty. Set the rank to 25 and run for 100 iterations. This corresponds to learning 25 topics. Plot the objective as a function of iteration. b) After running the algorithm, normalize the columns of W so they sum to one. For each column of W, list the 10 words having the largest weight and show the weight. The ith row of W corresponds to the ith word in the "dictionary" provided with the data. Organize these lists in a 5 x 5 table Comments about Problem 2: You can add a very small number (e.g., 10 16 to the denominator to avoid this. Each row in nyt-data .txt corresponds to a single document. It gives the index of words appearing in that document and the number of times they appear. It uses the format "idx:cnt" with commas separating each unique word in the document. Any index that doesn't appear in a row has a count of zero for that word. The vocabulary word corresponding to each index is given in the corresponding row of nyt vocab. dat