> For the complete documentation index, see [llms.txt](https://mugdha-thanawala.gitbook.io/tkmt_package/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://mugdha-thanawala.gitbook.io/tkmt_package/case-study/case-study-3.md).

# Case Study 3

The dataset was provided by the Prognostics CoE at NASA Ames. It is of a Li-ion battery which was run through 3 different operational profiles (charge, discharge and impedance) at room temperature. Here, the discharge capacity is the output feature while temperature, current and voltage are all input features. 

Importing the libraries 

```python
from tkmt_package.preprocessing import Preprocessing
from tkmt_package.ensemble import Rank_weighted
from tkmt_package.evaluation import Evaluation
import pandas as pd
from sklearn.ensemble import RandomForestRegressor
from sklearn.neighbors import KNeighborsRegressor
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Ridge
from sklearn.ensemble import ExtraTreesRegressor
```

Importing the dataset

```python
df = pd.read_csv(r"https://raw.githubusercontent.com/tkmtpackage/tkmt_ensemble/main/B0005.csv")
df.head()
```

![](/files/we7v6oeEAdNxMQZfXEpJ)

Rearranging the features

```python
dy= df[['ambient_temperature','voltage_measured', 'current_measured', 'temperature_measured',
       'current_load', 'voltage_load', 'capacity']]
```

Data Preprocessing

```python
prep= Preprocessing() 
X1,y1= prep.data_preparation(dy)
X_scal,y_scal = prep.data_normalization()
X_scal.shape,y_scal.shape
x_train,x_test,y_train,y_test = prep.data_split_train_test(X_scal,y_scal)
```

Rank Weighted Averaging

```python
rkwt= Rank_weighted()
base_model =[('rfr', RandomForestRegressor()), ('knn', KNeighborsRegressor()), ('lr', LinearRegression()),
             ('rd', Ridge()),('etr', ExtraTreesRegressor()) ]
base_model
summary = rkwt.get_weights(threshold=0.5,
    base_model = base_model,
    train_X = x_train,
    test_X  = x_test,
    train_y = y_train,
    test_y  = y_test)
```

![](/files/Ouhmn2atHz4in85ClZiM)

```python
modified_base_model =[('etr', ExtraTreesRegressor()), ('rfr', RandomForestRegressor()), ('knn', KNeighborsRegressor())]
modified_base_model
y_pred3 = rkwt.get_rank_weighted_technique(base_model=modified_base_model,
    train_X = x_train,
    train_y = y_train,
    test_X  = x_test,
    weights=[0.500000, 0.333333,0.166667])
```

Performance Evaluation

```python
eva = Evaluation()
eva.performance_evaluation(y_test,y_pred3)
```

![](/files/Qg8st6kRDjC7goVFYxwk)

```python
true2,pred2 =eva.get_plot_regression(true=y_test, pred=y_pred3, ascending=False, label='Discharge Capacity')
```

![](/files/eXkDFPbKyJebLUp73n34)

```python
# convert scaled data into its original format.
import numpy as np
true2 = prep.scaled_y.inverse_transform(np.array(y_test).reshape(-1,1))
pred2 =  prep.scaled_y.inverse_transform(np.array(y_pred3).reshape(-1,1))
true0,pred0 =eva.get_plot_regression(true=true2, pred=pred2, ascending=False, label='Discharge Capacity')
```

![](/files/WQoQe4T2eSU2jvA6NDOV)