TypeError: can't multiply sequence by non-int of type 'float'

[anfederico]

TypeError: can't multiply sequence by non-int of type 'float'

[anfederico]

Can I see the first 5 lines of your data file?

[vinodpkd]

Can I see the first 5 lines of your data file?

Here is the data format:

Date,Open,High,Low,Close,Volume
28-Oct-16,53.65,53.84,53.11,53.53,6620333
27-Oct-16,53.60,53.83,53.13,53.59,7899957
26-Oct-16,53.60,53.84,53.36,53.63,5817798
25-Oct-16,54.10,54.17,53.50,53.67,6052830
24-Oct-16,53.90,54.46,53.89,54.18,6919714

[anfederico]

One thing that may be an issue is your dates should be in the opposite order. Also, I don't see the variables that you listed in the header of your csv file.

variables = ["SSO", "SSC"] # If you set "SSO" and "SSC", they should be in your header

Try to format your data so it looks like this

Date,Open,High,Low,Close,Volume,SSO,SCC
03/01/2013,27.72,27.98,27.52,27.95,34851872,65.7894736842,-0.121
03/04/2013,27.85,28.15,27.7,28.15,38167504,75.9450171821,0.832
03/05/2013,28.29,28.54,28.16,28.35,41437136,84.9230769231,0.151
03/06/2013,28.21,28.23,27.78,28.09,51448912,80.7799442897,-0.689
03/07/2013,28.11,28.28,28.005,28.14,29197632,73.5368956743,-0.821

[vinodpkd]

Can I make
variables = []
instead of
variables = ["SSO", "SSC"]
I don't understand SSO and SSC.
I downloaded the data from google finance.

[vinodpkd]

data.iloc[0]
Out[20]:
Date 01/02/2013
Open 27.3
High 27.5
Low 27.13
Close 27.5
Volume 13268930
Name: 0, dtype: object

data.iloc[0].round(3)
Traceback (most recent call last):

File "", line 1, in
data.iloc[0].round(3)

File "C:\Anaconda2\lib\site-packages\pandas\core\series.py", line 1234, in round
result = _values_from_object(self).round(decimals, out=out)

TypeError: can't multiply sequence by non-int of type 'float'

data = data.round(3) # Round all values
is causing the error. Might it cannot round the date?

[uclatommy]

vinodpkd, SSO and SSC are custom-designed indicators that the author is using to train his model. At the end of his readme he provides a link to another project that is producing these values. Looks like he's mining social media to provide a social sentiment score.

If you don't understand the values, I think you can use any other data to train your model. For example, p/e ratios if you're more familiar with that indicator. Basically this training data is essential for the model to "learn" how to indicate a buy or sell. It uses a probabilistic classifier to over a training set in order to develop an association between the indicators and a buy or sell recommendation. So I think before you can continue any further, you have to provide at least some data as the indicator.

[anfederico]

Exactly, for example I see your dataset contains the header:
"Date,Open,High,Low,Close,Volume"

You could then do something like variables = ["High", "Low"] and the program will attempt to learn when to buy and when to sell based on the High and Low values of your data. Note that this is probably not going to work because the raw High/Low values alone are not very predictive of how the stock price will move. Therefore, you should try different indicators, like P/E for example.

"Date,Open,High,Low,Close,Volume,P/E"

Hope this helps!