Gold ETF Price Prediction using the Bayesian Ridge Linear Regression

Featured Photo by Pixabay.

• Yesterday @Barchart shared the Gold’s performance chart against the S&P 500 after Yield Curve Inversions. Is the uptrend for safe-havens set to continue?

Gold's performance against the S&P 500 after Yield Curve Inversions. Is Gold just getting started? 👀 pic.twitter.com/dSiLcBEbGb

— Barchart (@Barchart) March 23, 2023

• According to Florian Kössler, you would expect Gold to be at 1000$

Junior Gold & Silver Miners are close to an all time low compared to Majors.

Looking at this ratio, you would expect Gold to be at 1000$.

Should see a major shift here within the next 12-16 month.#GDX #GDXJ pic.twitter.com/sARduXzvYu

— Florian Kössler (@studentofcycles) March 23, 2023

• Thorsten Polleit also believes that it is time to buy Gold

My recommendation:
Sell your fiat money, buy physical gold. pic.twitter.com/4fl9kQ8IJw

— Thorsten Polleit (@ThorstenPolleit) March 22, 2023

• The objective of this post is to predict Gold price using Machine Learning (ML) algorithms in Python.

• Following the step-by-step guide, we will create a Machine Learning linear regression model that takes information from the past Gold ETF (GLD) prices and returns a Gold price prediction the next day.

• Recall that GLD is the largest ETF to invest directly in physical gold.

Let’s set the working directory GOLD

import os
os.chdir(‘GOLD’)
os. getcwd()

and import the following libraries

from sklearn.linear_model import LinearRegression

import pandas as pd
import numpy as np

import matplotlib.pyplot as plt
%matplotlib inline
plt.style.use(‘seaborn-darkgrid’)

import yfinance as yf

Let’s read the data
Df = yf.download(‘GLD’, ‘2022-01-01’, ‘2023-03-25’, auto_adjust=True)

Df = Df[[‘Close’]]

Df = Df.dropna()

Let’s plot the closing price of GLD
Df.Close.plot(figsize=(10, 7),color=’r’)
plt.ylabel(“Gold ETF Prices”)
plt.title(“Gold ETF Price Series”)

plt.savefig(‘goldetfprice.png’)

Let’s define the explanatory variables

Df[‘S_3’] = Df[‘Close’].rolling(window=3).mean()
Df[‘S_9’] = Df[‘Close’].rolling(window=9).mean()
Df[‘next_day_price’] = Df[‘Close’].shift(-1)

Df = Df.dropna()
X = Df[[‘S_3’, ‘S_9’]]

and the target variable

y = Df[‘next_day_price’]

Let’s split the data into the train and test datasets
t = .8
t = int(t*len(Df))

X_train = X[:t]
y_train = y[:t]

X_test = X[t:]
y_test = y[t:]

Let’s create the following linear regression model

from sklearn import linear_model
clf = linear_model.BayesianRidge()
linear = clf.fit(X_train, y_train)

Predicting the Gold ETF prices
predicted_price = linear.predict(X_test)
predicted_price = pd.DataFrame(
predicted_price, index=y_test.index, columns=[‘price’])
predicted_price.plot(figsize=(10, 7))
y_test.plot()
plt.legend([‘predicted_price’, ‘actual_price’])
plt.ylabel(“Gold ETF Price”)

plt.savefig(‘goldetfpricepredicted.png’)

The R2-score is
r2_score = linear.score(X[t:], y[t:])*100
float(“{0:.2f}”.format(r2_score))

79.81

Let’s plot the Cumulative Returns

gold = pd.DataFrame()

gold[‘price’] = Df[t:][‘Close’]
gold[‘predicted_price_next_day’] = predicted_price
gold[‘actual_price_next_day’] = y_test
gold[‘gold_returns’] = gold[‘price’].pct_change().shift(-1)

gold[‘signal’] = np.where(gold.predicted_price_next_day.shift(1) < gold.predicted_price_next_day,1,0)

gold[‘strategy_returns’] = gold.signal * gold[‘gold_returns’]
((gold[‘strategy_returns’]+1).cumprod()).plot(figsize=(10,7),color=’g’)
plt.ylabel(‘Cumulative Returns’)
plt.savefig(‘goldetfpricecumreturn.png’)

Let’s calculate the Sharpe ratio
sharpe = gold[‘strategy_returns’].mean()/gold[‘strategy_returns’].std()(252*0.5)
‘Sharpe Ratio %.2f’ % (sharpe)

'Sharpe Ratio 2.33'

Let’s get the forecast

import datetime as dt
current_date = dt.datetime.now()

data = yf.download(‘GLD’, ‘2022-01-01’, current_date, auto_adjust=True)
data[‘S_3’] = data[‘Close’].rolling(window=3).mean()
data[‘S_9’] = data[‘Close’].rolling(window=9).mean()
data = data.dropna()

data[‘predicted_gold_price’] = linear.predict(data[[‘S_3’, ‘S_9’]])
data[‘signal’] = np.where(data.predicted_gold_price.shift(1) < data.predicted_gold_price,”Buy”,”No Position”)

data.tail(1)[[‘signal’,’predicted_gold_price’]].T

Summary

• The SPDR Gold Trust Shares quote is equal to $185.220 at 2023-03-24, whereas our next-day forecast is $185.136.
• With the investment starting at 2022-12-15, the revenue is expected to be around +8%.
• The R2-score of our prediction model is about 80%.
• The Sharpe ratio of 2.33 is considered very good. The higher a fund’s Sharpe ratio, the better its returns have been relative to the amount of investment risk taken.
• Results fully support the Barchart opinion – 100% BUY Overall Average Signal calculated from all 13 indicators.

Explore More

Towards Max(ROI/Risk) Trading in Q1 2023

SARIMAX X-Validation of EIA Crude Oil Prices Forecast in 2023 – 2. Brent

SARIMAX X-Validation of EIA Crude Oil Prices Forecast in 2023 – 1. WTI

Stock Market ’22 Round Up & ’23 Outlook: Zacks Strategy vs Seeking Alpha Tactics

XOM SMA-EMA-RSI Golden Crosses ’22

BTC-USD Freefall vs FB/Meta Prophet 2022-23 Predictions

BTC-USD Price Prediction with LSTM Keras

All Eyes on ETFs Sep ’22

Towards min(Risk/Reward) – SeekingAlpha August Bear Market Update

ML/AI Regression for Stock Prediction – AAPL Use Case

---