Error in 10-fold cross validation code in Python

Question

I was implementing 10-fold cross validation from scratch in Python. The language is Python 3.6 and I wrote this in Spyder (Anaconda). My input shape is data=(1440,390),label=(1440,1).

My code:

def partitions(X,y):
  np.random.shuffle(X)
  foldx=[]
  foldy=[]
  j=0
  for i in range(0,10):
    foldx[i]=X[j:j+143,:]
    foldy[i]=y[foldx[j]]
    j+=144
  return np.array(foldx),np.array(foldy)

def cv(X,y,model):
  trainx,trainy=partitions(X,y)
  scores=[]
  for i in range(0,10):
    xtest=trainx[i]
    ytest=trainy[xtest]
    xtrain=trainx[:i]+trainx[i+1:]
    ytrain=trainy[xtrain]
    model.fit(xtrain,ytrain)
    preds=model.predict(xtest)
    print(accuracy_score(np.ravel(ytest),preds))
    scores.append(accuracy_score(np.ravel(ytest),preds))
  return scores.mean()

The error comes at

foldx[i]=X[j:j+143,:]

where it says

IndexError: list assignment index out of range.

How do I rectify this? I am not very experienced in implementing such problems from scratch.

Answer 1

You have to first populate a list to use it's indices, change the foldx[i]=X[j:j+143,:] line to

 foldx.append(X[j:j+143,:])

Similarly for foldy

foldy.append(y[foldx[j]])