Predictions using python GEKKO

Question

I have created an ARX model using sysid function, and now I want to extrapolate this model for other inputs.

I firstly created this model and identified its parameters for a training subset of data.

m = GEKKO(remote=False)
na = n_a
nb = n_b
yp_train, p, K = m.sysid(t_train, u_train, y_train, na, nb)

Having this model fit is there any way i can use the same parameters and utilize them with different inputs ?

Here is what i was going for

def arxouriço2(df, inputs, outputs, n_a, n_b, train_slice=None, test_slice=None):
    datafreime_arx = df.copy()
    if isinstance(inputs, pd.Series):
        inputs = pd.DataFrame(inputs)
    if isinstance(outputs, pd.Series):
        outputs = pd.DataFrame(outputs)
    colunas_arx = list(inputs.columns) + list(outputs.columns)
    mean_col = []
    std_dev_col = []

    for i in colunas_arx:
        mean = np.nanmean(datafreime_arx[i])
        mean_col.append(mean)
        datafreime_arx[i] -= mean

        std_dev = np.nanstd(datafreime_arx[i])
        std_dev_col.append(std_dev)
        datafreime_arx[i] /= std_dev

    if train_slice is not None:
        train_data = datafreime_arx.iloc[train_slice]
        t_train = train_data['Tempo']
        u_train = inputs.iloc[train_slice]
        y_train = outputs.iloc[train_slice]
    else:
        t_train = datafreime_arx['Tempo']
        u_train = inputs
        y_train = outputs

    if test_slice is not None:
        test_data = datafreime_arx.iloc[test_slice]
        t_test = test_data['Tempo']
        u_test = inputs.iloc[test_slice]
        y_test = outputs.iloc[test_slice]
    else:
        t_test = datafreime_arx['Tempo']
        u_test = inputs
        y_test = outputs

    m = GEKKO(remote=False)
    na = n_a
    nb = n_b
    yp_train, p, K = m.sysid(t_train, u_train, y_train, na, nb)

    for i, col in enumerate(colunas_arx):
        datafreime_arx[col] /= std_dev_col[i]
        datafreime_arx[col] -= mean_col[i]

    resis = y_train - yp_train
    resis2 = resis ** 2
    print('Sum of squared residuals: ' + str(np.sum(resis2)))

    if test_slice is not None:
        yp_test = m.sysid(t_test, u_test, y_test, na, nb)[0]
        plt.figure(figsize=(8, 5))
        plt.subplot(2, 1, 1)
        plt.plot(t_train, u_train)
        plt.legend(inputs.columns)
        plt.ylabel('Temperature (ºC)')
        plt.subplot(2, 1, 2)
        plt.plot(t_test, y_test, label=outputs.columns)
        plt.plot(t_test, yp_test, label=outputs.columns + ' predicted')
        plt.legend()
        plt.ylabel('Rotation (º)')
        plt.xlabel('Time')
        plt.tight_layout()
        plt.show()

    return p, K, m

arxouriço2(df_c3_avg5_c2,df_c3_avg5_c2[['STH3_avg','STH-04']],df_c3_avg5_c2['C14-Y'],10,10,range(465, 829),range(829, 903))

John Hedengren · Accepted Answer · 2024-04-12 17:50:35Z

1

Use the m.arx() function to create a model for predictions. The m.sysid() function is to determine the parameters p of an ARX model. The m.arx() function uses p parameters to create predictions. Here is an example with data:

Fit ARX with sysid

from gekko import GEKKO
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# load data and parse into columns
url = 'http://apmonitor.com/pdc/uploads/Main/'
data = pd.read_csv(url+'tclab_data3.txt')
t = data['Time']
u = data['Q1']
y = data['T1']

# generate time-series model
m = GEKKO(remote=False)

# system identification
na = 2 # output coefficients
nb = 2 # input coefficients
yp,p,K = m.sysid(t,u,y,na,nb,pred='meas')

plt.figure(figsize=(10,6))
plt.subplot(2,1,1)
plt.plot(t,u)
plt.legend([r'$Q_1$ (%)'])
plt.ylabel('MV Heater (%)')
plt.subplot(2,1,2)
plt.plot(t,y,'b-',label=r'$T_{1,meas}$')
plt.plot(t,yp,'r--',label=r'$T_{1,pred}$')
plt.legend(); plt.ylabel('CV Temp (°C)')
plt.xlabel('Time (sec)')

Predictions with ARX model using arx function

# load test data
data = pd.read_csv(url+'tclab_data4.txt')
t = data['Time']
um = data['Q1']
ym = data['T1']

# generate time-series model
m = GEKKO(remote=False)

# Build GEKKO ARX model
y,u = m.arx(p)
y[0].value = ym[0]

# load inputs
m.time = t
u[0].value = um

# options
m.options.imode = 4
m.options.nodes = 2

# simulate
m.solve(disp=False)

plt.figure(figsize=(10,6))
plt.subplot(2,1,1)
plt.plot(t,u[0])
plt.legend([r'$Q_1$ (%)'])
plt.ylabel('MV Heater (%)')
plt.subplot(2,1,2)
plt.plot(t,ym,'b-',label=r'$T_{1,meas}$')
plt.plot(t,y[0],'r--',label=r'$T_{1,pred}$')
plt.legend(); plt.ylabel('CV Temp (°C)')
plt.xlabel('Time (sec)')

For your function, modify this section of code:

if test_slice is not None:
    yp_test = m.sysid(t_test, u_test, y_test, na, nb)[0]

To something like this:

if test_slice is not None:
    mt = GEKKO(remote=False)
    y,u = mt.arx(p)
    mt.time = t_test
    y[0].value = y_test[0] # initial condition
    u[0].value = u_test
    mt.options.IMODE = 4
    mt.options.NODES = 2
    mt.solve(disp=False)
    yp_test = y[0].value

answered Apr 12, 2024 at 17:50

John Hedengren

14.6k1 gold badge24 silver badges32 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

yoitsBenjamin Over a year ago

Firstly thank you for the answer. I have two questions. 1. When I was first trying to fit a model I was able to use two columns as inputs and it worked fine, but now I can't. 2. The results I'm getting fitting the model are strange. The peaks are inverted and when solving the model for the testing subset there's a big descending stem in the begging that from my seeing doesn't make much sense. The testing subset starts right after the ending of the training one. imgur.com/a/0aq1Wqm

John Hedengren Over a year ago

Could you create a new question with complete code and data so that we can reproduce it? What version of gekko are you using? You can determine the version with “pip list”

Collectives™ on Stack Overflow

Predictions using python GEKKO

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related