def run_strategy(start_date, end_date):
print('')
#ms = manualStrategy(symbol = 'JPM', sd=start_date, ed=end_date, sv = 100000)
np.random.seed(106347)
threshold_list = [0.005, 0.01, 0.02, 0.03, 0.04, 0.05, 0.10]
for i in threshold_list:
learner = StrategyLearner(t=i)
learner.addEvidence(symbol = "JPM", sd=dt.datetime(2008,1,1), ed=dt.datetime(2009,12,31), sv = 100000)
df_trades = learner.testPolicy(symbol = "JPM", sd=start_date, ed=end_date, sv = 100000)
df_orders = learner.createOrdersDataFrame(df_trades)
portvals_strategy = msim.compute_portvals(orders_file = df_orders, start_val = 100000, commission=0, impact=0.005, end_date = end_date, start_date = start_date)
if i == threshold_list[0]:
threshold_portvals = pd.DataFrame(index = df_trades.index)
threshold_portvals[i] = portvals_strategy
print('Statistics for Random Forest Strategy with Buy/Sell Threshold: ' + str(i))
msim.computeStatistics(portvals_strategy)
import matplotlib.pyplot as plt
ax = threshold_portvals.plot(title='Portfolio Performance with different values of Buy/Sell Threshold', fontsize=10)
ax.set_xlabel('Date')
ax.set_ylabel('Portfolio Value')
plt.show()
def run_strategy(start_date, end_date):
print('')
ms = manualStrategy(symbol='JPM',
sd=start_date,
ed=end_date,
sv=100000)
df_trades = ms.testPolicy()
df_orders = ms.createOrdersDataFrame(df_trades)
benchmark = ms.createBenchMarkDataFrame(df_trades, sd=start_date)
portvals = msim.compute_portvals(orders_file=df_orders,
start_val=100000,
commission=9.95,
impact=0.005,
end_date=end_date,
start_date=start_date)
portvals_benchmark = msim.compute_portvals(orders_file=benchmark,
start_val=100000,
commission=0,
impact=0,
end_date=end_date,
start_date=start_date)
print('Statistics for Manual Strategy')
msim.computeStatistics(portvals)
print('')
print('Statistics for Benchmark Strategy')
msim.computeStatistics(portvals_benchmark)
generatePerformanceChart(df_trades, portvals, portvals_benchmark)
def run_strategy(start_date, end_date):
print('')
#ms = manualStrategy(symbol = 'JPM', sd=start_date, ed=end_date, sv = 100000)
np.random.seed(106347)
learner = StrategyLearner()
learner.addEvidence(symbol="JPM",
sd=dt.datetime(2008, 1, 1),
ed=dt.datetime(2009, 12, 31),
sv=100000)
df_trades = learner.testPolicy(symbol="JPM",
sd=start_date,
ed=end_date,
sv=100000)
#df_trades = ms.testPolicy()
df_orders = learner.createOrdersDataFrame(df_trades)
benchmark = learner.createBenchMarkDataFrame(df_trades, sd=start_date)
portvals = msim.compute_portvals(orders_file=df_orders,
start_val=100000,
commission=0,
impact=0.005,
end_date=end_date,
start_date=start_date)
portvals_benchmark = msim.compute_portvals(orders_file=benchmark,
start_val=100000,
commission=0,
impact=0,
end_date=end_date,
start_date=start_date)
print 'Statistics for Random Forest Strategy'
msim.computeStatistics(portvals)
print ''
print 'Statistics for Benchmark Strategy'
msim.computeStatistics(portvals_benchmark)
generatePerformanceChart(df_trades, portvals, portvals_benchmark)
def run_strategy(start_date, end_date):
print('')
#ms = manualStrategy(symbol = 'JPM', sd=start_date, ed=end_date, sv = 100000)
np.random.seed(106348)
symbols = ['AAPL', 'XOM', 'GOOG', 'IBM']
for sym in symbols:
print sym
learner = StrategyLearner()
learner.addEvidence(symbol=sym,
sd=dt.datetime(2008, 1, 1),
ed=dt.datetime(2009, 12, 31),
sv=100000)
df_trades = learner.testPolicy(symbol=sym,
sd=start_date,
ed=end_date,
sv=100000)
df_orders = learner.createOrdersDataFrame(df_trades)
ms = manual.manualStrategy(symbol=sym,
sd=start_date,
ed=end_date,
sv=100000)
df_trades_ms = ms.testPolicy()
df_orders_ms = ms.createOrdersDataFrame(df_trades_ms)
benchmark = learner.createBenchMarkDataFrame(df_trades,
sd=start_date)
portvals_strategy = msim.compute_portvals(orders_file=df_orders,
start_val=100000,
commission=0,
impact=0.005,
end_date=end_date,
start_date=start_date)
portvals_benchmark = msim.compute_portvals(orders_file=benchmark,
start_val=100000,
commission=0,
impact=0,
end_date=end_date,
start_date=start_date)
portvals_manual = msim.compute_portvals(orders_file=df_orders_ms,
start_val=100000,
commission=0,
impact=0.005,
end_date=end_date,
start_date=start_date)
print('Statistics for Random Forest Strategy')
msim.computeStatistics(portvals_strategy)
print('')
print('Statistics for Manual Strategy')
msim.computeStatistics(portvals_manual)
print('')
print('Statistics for Benchmark Strategy')
msim.computeStatistics(portvals_benchmark)
generatePerformanceChart(df_trades, portvals_strategy,
portvals_manual, portvals_benchmark)
def run_strategy(start_date, end_date):
print('')
#ms = manualStrategy(symbol = 'JPM', sd=start_date, ed=end_date, sv = 100000)
np.random.seed(106348)
learner = StrategyLearner()
learner.addEvidence(symbol="JPM",
sd=dt.datetime(2008, 1, 1),
ed=dt.datetime(2009, 12, 31),
sv=100000)
df_trades = learner.testPolicy(symbol="JPM",
sd=start_date,
ed=end_date,
sv=100000)
df_orders = learner.createOrdersDataFrame(df_trades)
ms = manual.manualStrategy(symbol='JPM',
sd=start_date,
ed=end_date,
sv=100000)
df_trades_ms = ms.testPolicy()
df_orders_ms = ms.createOrdersDataFrame(df_trades_ms)
benchmark = learner.createBenchMarkDataFrame(df_trades, sd=start_date)
portvals_strategy = msim.compute_portvals(orders_file=df_orders,
start_val=100000,
commission=0,
impact=0.005,
end_date=end_date,
start_date=start_date)
portvals_benchmark = msim.compute_portvals(orders_file=benchmark,
start_val=100000,
commission=0,
impact=0,
end_date=end_date,
start_date=start_date)
portvals_manual = msim.compute_portvals(orders_file=df_orders_ms,
start_val=100000,
commission=0,
impact=0.005,
end_date=end_date,
start_date=start_date)
print('Statistics for Random Forest Strategy')
msim.computeStatistics(portvals_strategy)
print('')
print('Statistics for Manual Strategy')
msim.computeStatistics(portvals_manual)
print('')
print('Statistics for Benchmark Strategy')
msim.computeStatistics(portvals_benchmark)
generatePerformanceChart(df_trades, portvals_strategy, portvals_manual,
portvals_benchmark)
tot_cum_ret = 0
tot_sharpe_ratio = 0
tot_std_daily_ret = 0
tot_avg_daily_ret = 0
N = 10
i = 0
print ''
print 'Lets run the Random Forest 10 times and average results for performance'
while i < N:
np.random.seed(np.random.randint(0, 100000))
learner = StrategyLearner()
learner.addEvidence(symbol="JPM",
sd=dt.datetime(2008, 1, 1),
ed=dt.datetime(2009, 12, 31),
sv=100000)
df_trades = learner.testPolicy(symbol="JPM",
sd=start_date,
ed=end_date,
sv=100000)
df_orders = learner.createOrdersDataFrame(df_trades)
portvals_strategy = msim.compute_portvals(orders_file=df_orders,
start_val=100000,
commission=0,
impact=0.005,
end_date=end_date,
start_date=start_date)
cum_ret, sharpe_ratio, std_daily_ret, avg_daily_ret = msim.computeStatisticsReturnValues(
portvals_strategy)
tot_cum_ret = tot_cum_ret + cum_ret
tot_sharpe_ratio = tot_sharpe_ratio + sharpe_ratio
tot_std_daily_ret = tot_std_daily_ret + std_daily_ret
i = i + 1
print 'The average cumulative return is: ' + str(tot_cum_ret / N)
print 'The average sharpe ratio is: ' + str(tot_sharpe_ratio / N)
print 'The average standard dev of daily returns is: ' + str(
tot_std_daily_ret / N)