kartik-joshi/stock-predection

Stock Prediction using machine learning

stock-predection's Issues

About Error while Running SVM file in python

Traceback (most recent call last):
File "C:/Users/Broadway/PycharmProjects/pp1.py/SVM.py", line 48, in
testdataset =Read_file('APPL.csv')
File "C:/Users/Broadway/PycharmProjects/pp1.py/SVM.py", line 9, in Read_file
with open(file_name, 'r', newline='',encoding='utf-8') as file:
FileNotFoundError: [Errno 2] No such file or directory: 'APPL.csv'

Process finished with exit code 1

**how to solve this Error

doubt about the procedure

before applying all the algorithm you have done preprocessing by read file then converted string column into float then converted string into integer .
so my question is
is they can do it directly by using pandas function because i am getting error in that.

Output and prediction results

For your output, shouldn't the class be 1 if the stock price is higher the next day vs. today not if it is higher today vs. yesterday? If you are comparing today vs. yesterday, that is not a prediction.

Also, your prediction results are using the training sets as part of the prediction results, which are inflating the results; I've only reviewed the SVM code, but I assume the logic is the same in the other procedures.

For instance,

    for i in range(0, len(testdataset)):
        total += 1
        temp = clf.predict([testdataset[i]])
        if temp == test_target[i]:
            count += 1
    accuracy = count * 100 / total
    print('Accuracy: %s' % accuracy)
    runs[j] = accuracy

Should be,

    for i in range(r + 10000, len(testdataset)):
        total += 1
        temp = clf.predict([testdataset[i]])
        if temp == test_target[i]:
            count += 1
    accuracy = count * 100 / total
    print('Accuracy: %s' % accuracy)
    runs[j] = accuracy