Forex Forecasting with Random Forests 4/N – Precision Rating

### Trade Simulation
def trade_simulation(dfx,leverage,start_balance):
    i = 0
    for index,item in dfx.iterrows():
        # Margin required for 10,000 EUR (1 ticket) trade
        dfx.loc[index,'margin_per_eur'] = item['eurusd_Close']*item['usdjpy_Close']*10000/leverage

        # Bank balance
        if i <= 1:
            dfx.loc[index,'balance'] = start_balance 
        else:
            dfx.loc[index,'balance'] = dfx.loc[index-1,'balance'] + dfx.loc[index-1,'profit_loss']

        # Num
        num = dfx.loc[index,'balance']/2/dfx.loc[index,'margin_per_eur']
        if num>100:
            num = 100
        dfx.loc[index,'num']  = num

        # Required Margin
        dfx.loc[index,'margin']  = dfx.loc[index,'margin_per_eur']*dfx.loc[index,'num']

        # Investment Residuals after Position Making
        dfx.loc[index,'capacity']  = dfx.loc[index,'balance']-dfx.loc[index,'margin']

        # Rate fluctuation tolerance
        dfx.loc[index,'rate_fluctuation_tolerance'] = dfx.loc[index,'capacity']/(dfx.loc[index,'num']*10000)

        # Profit and loss
        if i <= 1:
            dfx.loc[index,'profit_loss'] = 0 
        else:
            dfx.loc[index,'profit_loss'] = dfx.loc[index-2,'pred']*(dfx.loc[index,'eurusd_Close']-dfx.loc[index-2,'eurusd_Close'])*dfx.loc[index-2,'num']*10000*dfx.loc[index,'usdjpy_Close']

        i = i+1
    return dfx

### Trade Simulation
def trade_simulation(dfx,leverage,start_balance):
    i = 0
    for index,item in dfx.iterrows():
        # Margin required for 10,000 EUR (1 ticket) trade
        dfx.loc[index,'margin_per_eur'] = item['eurusd_Close']*item['usdjpy_Close']*10000/leverage

        # Bank balance
        if i <= 1:
            dfx.loc[index,'balance'] = start_balance 
        else:
            dfx.loc[index,'balance'] = dfx.loc[index-1,'balance'] + dfx.loc[index-1,'profit_loss']

        # Num
        num = dfx.loc[index,'balance']/2/dfx.loc[index,'margin_per_eur']
        if num>100:
            num = 100
        dfx.loc[index,'num']  = num

        # Required Margin
        dfx.loc[index,'margin']  = dfx.loc[index,'margin_per_eur']*dfx.loc[index,'num']

        # Investment Residuals after Position Making
        dfx.loc[index,'capacity']  = dfx.loc[index,'balance']-dfx.loc[index,'margin']

        # Rate fluctuation tolerance
        dfx.loc[index,'rate_fluctuation_tolerance'] = dfx.loc[index,'capacity']/(dfx.loc[index,'num']*10000)

        # Profit and loss
        if i <= 1:
            dfx.loc[index,'profit_loss'] = 0 
        else:
            dfx.loc[index,'profit_loss'] = dfx.loc[index-2,'pred']*(dfx.loc[index,'eurusd_Close']-dfx.loc[index-2,'eurusd_Close'])*dfx.loc[index-2,'num']*10000*dfx.loc[index,'usdjpy_Close']

        i = i+1
    return dfx

### Trade Simulation
def trade_simulation(dfx,leverage,start_balance):
    i = 0
    for index,item in dfx.iterrows():
        # Margin required for 10,000 EUR (1 ticket) trade
        dfx.loc[index,'margin_per_eur'] = item['eurusd_Close']*item['usdjpy_Close']*10000/leverage

        # Bank balance
        if i <= 1:
            dfx.loc[index,'balance'] = start_balance 
        else:
            dfx.loc[index,'balance'] = dfx.loc[index-1,'balance'] + dfx.loc[index-1,'profit_loss']

        # Num
        num = dfx.loc[index,'balance']/2/dfx.loc[index,'margin_per_eur']
        if num>100:
            num = 100
        dfx.loc[index,'num']  = num

        # Required Margin
        dfx.loc[index,'margin']  = dfx.loc[index,'margin_per_eur']*dfx.loc[index,'num']

        # Investment Residuals after Position Making
        dfx.loc[index,'capacity']  = dfx.loc[index,'balance']-dfx.loc[index,'margin']

        # Rate fluctuation tolerance
        dfx.loc[index,'rate_fluctuation_tolerance'] = dfx.loc[index,'capacity']/(dfx.loc[index,'num']*10000)

        # Profit and loss
        if i <= 1:
            dfx.loc[index,'profit_loss'] = 0 
        else:
            dfx.loc[index,'profit_loss'] = dfx.loc[index-2,'pred']*(dfx.loc[index,'eurusd_Close']-dfx.loc[index-2,'eurusd_Close'])*dfx.loc[index-2,'num']*10000*dfx.loc[index,'usdjpy_Close']

        i = i+1
    return dfx

## Combine explanatory variables with predicted results of test data
# Data framing of eries type data
test_y2 = pd.DataFrame(Y_test)
test_y2['pred'] = model.predict(X_test)

# Prediction Result Storage
col = ['Date','eurusd_Open', 'eurusd_High', 'eurusd_Low', 'eurusd_Close', 'usdjpy_Close', 'Maguro', 'pred']
df_test_result = pd.merge(df_drop_add, test_y2['pred'], left_index=True, right_index=True)[col]


## Combining explanatory variables with the predicted results of train data
# Data framing of Series type data
train_y2 = pd.DataFrame(Y_train)
train_y2['pred'] = model.predict(X_train)

# Prediction Result Saving
col = ['Date','eurusd_Open', 'eurusd_High', 'eurusd_Low', 'eurusd_Close', 'usdjpy_Close', 'Maguro', 'pred']
df_train_result = pd.merge(df_drop_add, train_y2['pred'], left_index=True, right_index=True)[col]

Ranking of articles this category