In [6]:

# !pip install yfinance # uncomment these to install missing packages if they are not already installed
# !pip install pandas 

import yfinance as yf
import pandas as pd

def get_price(tick,start='2022-10-01',end=None):
    return yf.Ticker(tick).history(start=start,end=end)['Close']

def get_prices(tickers,start='2022-10-01',end=None):
    df=pd.DataFrame()
    for s in tickers:
        df[s]=get_price(s,start,end)
    return df

Requirement already satisfied: yfinance in /home/phsamuel/p3/lib/python3.8/site-packages (0.1.55)
Requirement already satisfied: multitasking>=0.0.7 in /home/phsamuel/p3/lib/python3.8/site-packages (from yfinance) (0.0.9)
Requirement already satisfied: pandas>=0.24 in /home/phsamuel/p3/lib/python3.8/site-packages (from yfinance) (1.2.1)
Requirement already satisfied: requests>=2.20 in /home/phsamuel/p3/lib/python3.8/site-packages (from yfinance) (2.22.0)
Requirement already satisfied: lxml>=4.5.1 in /home/phsamuel/p3/lib/python3.8/site-packages (from yfinance) (4.6.2)
Requirement already satisfied: numpy>=1.15 in /home/phsamuel/p3/lib/python3.8/site-packages (from yfinance) (1.20.0)
Requirement already satisfied: pytz>=2017.3 in /usr/lib/python3/dist-packages (from pandas>=0.24->yfinance) (2019.3)
Requirement already satisfied: python-dateutil>=2.7.3 in /usr/lib/python3/dist-packages (from pandas>=0.24->yfinance) (2.7.3)
Requirement already satisfied: pandas in /home/phsamuel/p3/lib/python3.8/site-packages (1.2.1)
Requirement already satisfied: pytz>=2017.3 in /usr/lib/python3/dist-packages (from pandas) (2019.3)
Requirement already satisfied: numpy>=1.16.5 in /home/phsamuel/p3/lib/python3.8/site-packages (from pandas) (1.20.0)
Requirement already satisfied: python-dateutil>=2.7.3 in /usr/lib/python3/dist-packages (from pandas) (2.7.3)

Prepare training and testing data sets¶

In [2]:

feature_stocks=['tsla','meta','twtr','amzn','nflx','gbtc','gdx','intc','dal','c']
predict_stock='msft'

# training set
start_date_train='2022-10-01'
end_date_train='2022-12-31'

X_train=get_prices(feature_stocks,start=start_date_train,end=end_date_train)
y_train=get_prices([predict_stock],start=start_date_train,end=end_date_train)

# testing set
start_date_test='2023-01-01' # end date omit, default is doday
X_test=get_prices(feature_stocks,start=start_date_test)
y_test=get_prices([predict_stock],start=start_date_test)

In [4]:

X_train

Out[4]:

	tsla	fb	twtr	amzn	nflx	gbtc	gdx	intc	dal	c
Date
2020-10-01	448.160004	266.630005	46.700001	3221.260010	527.510010	10.870000	39.364471	51.862778	31.100000	42.545544
2020-10-02	415.089996	259.940002	46.119999	3125.000000	503.059998	10.860000	38.767590	50.641655	31.750000	42.761013
2020-10-05	425.679993	264.649994	47.310001	3199.199951	520.650024	11.280000	39.364471	51.316746	32.000000	43.985275
2020-10-06	413.980011	258.660004	45.599998	3099.959961	505.869995	10.845000	37.912056	50.999058	31.059999	43.495571
2020-10-07	425.299988	258.119995	45.869999	3195.689941	534.659973	10.970000	38.150806	52.289669	32.150002	43.916718
...	...	...	...	...	...	...	...	...	...	...
2020-12-23	645.979980	268.109985	54.299999	3185.270020	514.479980	28.879999	35.919998	46.570000	40.240002	60.266277
2020-12-24	661.770020	267.399994	53.970001	3172.689941	513.969971	27.350000	36.029999	47.070000	39.730000	60.058056
2020-12-28	663.690002	277.000000	54.430000	3283.959961	519.119995	30.450001	35.689999	47.070000	40.150002	60.613323
2020-12-29	665.989990	276.779999	54.360001	3322.000000	530.869995	30.080000	35.740002	49.389999	40.029999	60.395180
2020-12-30	694.780029	271.869995	54.330002	3285.850098	524.590027	32.900002	36.560001	48.750000	40.560001	60.345604

63 rows × 10 columns

In [5]:

y_train

Out[5]:

	msft
Date
2020-10-01	211.905228
2020-10-02	205.651596
2020-10-05	209.830658
2020-10-06	205.372330
2020-10-07	209.282089
...	...
2020-12-23	221.020004
2020-12-24	222.750000
2020-12-28	224.960007
2020-12-29	224.149994
2020-12-30	221.679993

63 rows × 1 columns

Convert training and testing data into numpy array¶

In [ ]:

import numpy as np

X_train=np.array(X_train)
y_train=np.array(y_train)
X_test=np.array(X_test)
y_test=np.array(y_test)

Use linear regression to predict msft stock price from the other stocks' prices¶

1. Append a dummy feature to both X_train and X_test¶

In [ ]:

# Your solution here

2. Find the best linear regression model based on your training data ($w=(X X')^{-1} X y$)¶

Note that you may need to transpose the matrices to make things work¶

In [ ]:

# Your solution here

3. Report your training and testing error¶

How far your prediction from the actual price. Compute the mean square error for both training and testing¶

In [ ]:

# Your solution here

In [ ]: