306 KiB
306 KiB
In [1]:
import pickle
import pandas as pd
from sklearn import tree
model = pickle.load(open("data/vtree.model.sav", "rb"))
features = (
pd.read_csv("data/viscosity_train.csv", sep=";", decimal=",")
.drop(["Viscosity"], axis=1)
.columns.values.tolist()
)
rules = tree.export_text(model, feature_names=features)
print(rules)
|--- T <= 32.50 | |--- TiO2 <= 0.18 | | |--- Al2O3 <= 0.18 | | | |--- T <= 22.50 | | | | |--- TiO2 <= 0.03 | | | | | |--- Al2O3 <= 0.03 | | | | | | |--- value: [3.71] | | | | | |--- Al2O3 > 0.03 | | | | | | |--- value: [4.66] | | | | |--- TiO2 > 0.03 | | | | | |--- value: [4.88] | | | |--- T > 22.50 | | | | |--- TiO2 <= 0.03 | | | | | |--- Al2O3 <= 0.03 | | | | | | |--- value: [3.18] | | | | | |--- Al2O3 > 0.03 | | | | | | |--- value: [3.38] | | | | |--- TiO2 > 0.03 | | | | | |--- value: [4.24] | | |--- Al2O3 > 0.18 | | | |--- T <= 22.50 | | | | |--- value: [6.67] | | | |--- T > 22.50 | | | | |--- T <= 27.50 | | | | | |--- value: [5.59] | | | | |--- T > 27.50 | | | | | |--- value: [4.73] | |--- TiO2 > 0.18 | | |--- T <= 22.50 | | | |--- value: [7.13] | | |--- T > 22.50 | | | |--- T <= 27.50 | | | | |--- value: [5.87] | | | |--- T > 27.50 | | | | |--- value: [4.94] |--- T > 32.50 | |--- T <= 47.50 | | |--- TiO2 <= 0.18 | | | |--- Al2O3 <= 0.18 | | | | |--- T <= 42.50 | | | | | |--- TiO2 <= 0.03 | | | | | | |--- Al2O3 <= 0.03 | | | | | | | |--- value: [2.36] | | | | | | |--- Al2O3 > 0.03 | | | | | | | |--- value: [2.68] | | | | | |--- TiO2 > 0.03 | | | | | | |--- T <= 37.50 | | | | | | | |--- value: [3.12] | | | | | | |--- T > 37.50 | | | | | | | |--- value: [2.65] | | | | |--- T > 42.50 | | | | | |--- TiO2 <= 0.03 | | | | | | |--- value: [1.83] | | | | | |--- TiO2 > 0.03 | | | | | | |--- value: [2.40] | | | |--- Al2O3 > 0.18 | | | | |--- T <= 37.50 | | | | | |--- value: [4.12] | | | | |--- T > 37.50 | | | | | |--- value: [3.56] | | |--- TiO2 > 0.18 | | | |--- T <= 40.00 | | | | |--- value: [4.35] | | | |--- T > 40.00 | | | | |--- value: [3.56] | |--- T > 47.50 | | |--- TiO2 <= 0.18 | | | |--- Al2O3 <= 0.18 | | | | |--- T <= 52.50 | | | | | |--- TiO2 <= 0.03 | | | | | | |--- Al2O3 <= 0.03 | | | | | | | |--- value: [1.63] | | | | | | |--- Al2O3 > 0.03 | | | | | | | |--- value: [1.90] | | | | | |--- TiO2 > 0.03 | | | | | | |--- value: [2.11] | | | | |--- T > 52.50 | | | | | |--- T <= 65.00 | | | | | | |--- TiO2 <= 0.03 | | | | | | | |--- value: [1.55] | | | | | | |--- TiO2 > 0.03 | | | | | | | |--- value: [1.66] | | | | | |--- T > 65.00 | | | | | | |--- TiO2 <= 0.03 | | | | | | | |--- value: [1.19] | | | | | | |--- TiO2 > 0.03 | | | | | | | |--- value: [1.29] | | | |--- Al2O3 > 0.18 | | | | |--- T <= 65.00 | | | | | |--- T <= 57.50 | | | | | | |--- value: [2.43] | | | | | |--- T > 57.50 | | | | | | |--- value: [2.16] | | | | |--- T > 65.00 | | | | | |--- value: [1.73] | | |--- TiO2 > 0.18 | | | |--- T <= 65.00 | | | | |--- T <= 57.50 | | | | | |--- value: [2.84] | | | | |--- T > 57.50 | | | | | |--- value: [2.54] | | | |--- T > 65.00 | | | | |--- value: [1.91]
In [2]:
from src.rules import get_rules
rules = get_rules(model, features)
display(len(rules))
rules
35
Out[2]:
[if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 52.5) and (T > 65.0) and (TiO2 <= 0.025) -> 1.194, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 52.5) and (T > 65.0) and (TiO2 > 0.025) -> 1.289, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 52.5) and (T <= 65.0) and (TiO2 <= 0.025) -> 1.548, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 52.5) and (TiO2 <= 0.025) and (Al2O3 <= 0.025) -> 1.629, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 52.5) and (T <= 65.0) and (TiO2 > 0.025) -> 1.662, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T > 65.0) -> 1.728, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 42.5) and (TiO2 <= 0.025) -> 1.832, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 52.5) and (TiO2 <= 0.025) and (Al2O3 > 0.025) -> 1.897, if (T > 32.5) and (T > 47.5) and (TiO2 > 0.175) and (T > 65.0) -> 1.91, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 52.5) and (TiO2 > 0.025) -> 2.109, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T <= 65.0) and (T > 57.5) -> 2.16, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 42.5) and (TiO2 <= 0.025) and (Al2O3 <= 0.025) -> 2.361, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 42.5) and (TiO2 > 0.025) -> 2.402, if (T > 32.5) and (T > 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T <= 65.0) and (T <= 57.5) -> 2.426, if (T > 32.5) and (T > 47.5) and (TiO2 > 0.175) and (T <= 65.0) and (T > 57.5) -> 2.538, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 42.5) and (TiO2 > 0.025) and (T > 37.5) -> 2.655, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 42.5) and (TiO2 <= 0.025) and (Al2O3 > 0.025) -> 2.682, if (T > 32.5) and (T > 47.5) and (TiO2 > 0.175) and (T <= 65.0) and (T <= 57.5) -> 2.838, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 42.5) and (TiO2 > 0.025) and (T <= 37.5) -> 3.121, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 22.5) and (TiO2 <= 0.025) and (Al2O3 <= 0.025) -> 3.18, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 22.5) and (TiO2 <= 0.025) and (Al2O3 > 0.025) -> 3.38, if (T > 32.5) and (T <= 47.5) and (TiO2 > 0.175) and (T > 40.0) -> 3.561, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T > 37.5) -> 3.565, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 22.5) and (TiO2 <= 0.025) and (Al2O3 <= 0.025) -> 3.707, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T <= 37.5) -> 4.118, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T > 22.5) and (TiO2 > 0.025) -> 4.236, if (T > 32.5) and (T <= 47.5) and (TiO2 > 0.175) and (T <= 40.0) -> 4.354, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 22.5) and (TiO2 <= 0.025) and (Al2O3 > 0.025) -> 4.66, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T > 22.5) and (T > 27.5) -> 4.731, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (T <= 22.5) and (TiO2 > 0.025) -> 4.885, if (T <= 32.5) and (TiO2 > 0.175) and (T > 22.5) and (T > 27.5) -> 4.944, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T > 22.5) and (T <= 27.5) -> 5.594, if (T <= 32.5) and (TiO2 > 0.175) and (T > 22.5) and (T <= 27.5) -> 5.865, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) and (T <= 22.5) -> 6.67, if (T <= 32.5) and (TiO2 > 0.175) and (T <= 22.5) -> 7.132]
In [3]:
from src.rules import normalise_rules
rules = normalise_rules(rules)
display(len(rules))
rules
35
Out[3]:
[if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.194, if (T > 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.289, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.548, if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.629, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.662, if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 1.728, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.832, if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 1.897, if (T > 32.5) and (TiO2 > 0.175) -> 1.91, if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.109, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 2.16, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 2.361, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.402, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 2.426, if (T > 32.5) and (T <= 65.0) and (TiO2 > 0.175) -> 2.538, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.655, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 2.682, if (T > 32.5) and (T <= 65.0) and (TiO2 > 0.175) -> 2.838, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 3.121, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 3.18, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 3.38, if (T > 32.5) and (T <= 47.5) and (TiO2 > 0.175) -> 3.561, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 3.565, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 3.707, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 4.118, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 4.236, if (T > 32.5) and (T <= 47.5) and (TiO2 > 0.175) -> 4.354, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 4.66, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 4.731, if (T <= 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 4.885, if (T <= 32.5) and (T > 22.5) and (TiO2 > 0.175) -> 4.944, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 5.594, if (T <= 32.5) and (T > 22.5) and (TiO2 > 0.175) -> 5.865, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 6.67, if (T <= 32.5) and (TiO2 > 0.175) -> 7.132]
In [4]:
from src.rules import delete_same_rules
rules = delete_same_rules(rules)
display(len(rules))
rules
26
Out[4]:
[if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.194, if (T > 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.289, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.548, if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.629, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.662, if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 1.728, if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 1.897, if (T > 32.5) and (TiO2 > 0.175) -> 1.91, if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.109, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 2.097, if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 2.293, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 2.682, if (T > 32.5) and (T <= 65.0) and (TiO2 > 0.175) -> 2.688, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.726, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 3.18, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 3.38, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 3.707, if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 3.842, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 4.236, if (T > 32.5) and (T <= 47.5) and (TiO2 > 0.175) -> 3.958, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 4.66, if (T <= 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 4.885, if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 5.162, if (T <= 32.5) and (T > 22.5) and (TiO2 > 0.175) -> 5.405, if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 6.67, if (T <= 32.5) and (TiO2 > 0.175) -> 7.132]
In [5]:
from src.rules import get_features, vectorize_rules
features = get_features(rules, ["T"])
print(features)
df_rules = vectorize_rules(rules, features)
df_rules.head(5)
['(Al2O3 <= 0.175)', '(Al2O3 > 0.025)', '(Al2O3 > 0.175)', '(TiO2 <= 0.175)', '(TiO2 > 0.025)', '(TiO2 > 0.175)']
Out[5]:
| (Al2O3 <= 0.175) | (Al2O3 > 0.025) | (Al2O3 > 0.175) | (TiO2 <= 0.175) | (TiO2 > 0.025) | (TiO2 > 0.175) | consequent | |
|---|---|---|---|---|---|---|---|
| rule | |||||||
| if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.194 | 1 | 0 | 0 | 1 | 0 | 0 | 1.194 |
| if (T > 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.289 | 1 | 0 | 0 | 1 | 1 | 0 | 1.289 |
| if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.548 | 1 | 0 | 0 | 1 | 0 | 0 | 1.548 |
| if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.629 | 1 | 0 | 0 | 1 | 0 | 0 | 1.629 |
| if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.662 | 1 | 0 | 0 | 1 | 1 | 0 | 1.662 |
In [6]:
from src.cluster_helper import draw_best_clusters_plot, get_best_clusters_num
random_state = 9
X = df_rules.copy()
X = X.drop(["consequent"], axis=1)
clusters_score = get_best_clusters_num(X, random_state)
display(clusters_score)
draw_best_clusters_plot(clusters_score)
clusters_num = sorted(clusters_score.items(), key=lambda x: x[1], reverse=True)[0][0]
display(f"The best clusters count is {clusters_num}")
c:\Users\user\Projects\python\fuzzy\.venv\Lib\site-packages\sklearn\base.py:1473: ConvergenceWarning: Number of distinct clusters (5) found smaller than n_clusters (6). Possibly due to duplicate points in X. return fit_method(estimator, *args, **kwargs) c:\Users\user\Projects\python\fuzzy\.venv\Lib\site-packages\sklearn\base.py:1473: ConvergenceWarning: Number of distinct clusters (5) found smaller than n_clusters (7). Possibly due to duplicate points in X. return fit_method(estimator, *args, **kwargs) c:\Users\user\Projects\python\fuzzy\.venv\Lib\site-packages\sklearn\base.py:1473: ConvergenceWarning: Number of distinct clusters (5) found smaller than n_clusters (8). Possibly due to duplicate points in X. return fit_method(estimator, *args, **kwargs) c:\Users\user\Projects\python\fuzzy\.venv\Lib\site-packages\sklearn\base.py:1473: ConvergenceWarning: Number of distinct clusters (5) found smaller than n_clusters (9). Possibly due to duplicate points in X. return fit_method(estimator, *args, **kwargs) c:\Users\user\Projects\python\fuzzy\.venv\Lib\site-packages\sklearn\base.py:1473: ConvergenceWarning: Number of distinct clusters (5) found smaller than n_clusters (10). Possibly due to duplicate points in X. return fit_method(estimator, *args, **kwargs)
{2: 0.42601604060235754,
3: 0.7018993361356377,
4: 0.8249121250065079,
5: 1.0,
6: 1.0,
7: 1.0,
8: 1.0,
9: 1.0,
10: 1.0}
'The best clusters count is 5'
In [7]:
from sklearn import cluster
from src.cluster_helper import print_cluster_result
kmeans = cluster.KMeans(n_clusters=clusters_num, random_state=random_state)
kmeans.fit(X)
print_cluster_result(X, clusters_num, kmeans.labels_)
Кластер 1 (6): if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.194; if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.548; if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 1.629; if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 2.097; if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 3.18; if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) -> 3.707 -------- Кластер 2 (5): if (T > 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 1.728; if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 2.293; if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 3.842; if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 5.162; if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 > 0.175) -> 6.67 -------- Кластер 3 (5): if (T > 32.5) and (TiO2 > 0.175) -> 1.91; if (T > 32.5) and (T <= 65.0) and (TiO2 > 0.175) -> 2.688; if (T > 32.5) and (T <= 47.5) and (TiO2 > 0.175) -> 3.958; if (T <= 32.5) and (T > 22.5) and (TiO2 > 0.175) -> 5.405; if (T <= 32.5) and (TiO2 > 0.175) -> 7.132 -------- Кластер 4 (6): if (T > 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.289; if (T > 32.5) and (T <= 65.0) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 1.662; if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.109; if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 2.726; if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 4.236; if (T <= 32.5) and (TiO2 <= 0.175) and (TiO2 > 0.025) and (Al2O3 <= 0.175) -> 4.885 -------- Кластер 5 (4): if (T > 32.5) and (T <= 52.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 1.897; if (T > 32.5) and (T <= 47.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 2.682; if (T <= 32.5) and (T > 22.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 3.38; if (T <= 32.5) and (TiO2 <= 0.175) and (Al2O3 <= 0.175) and (Al2O3 > 0.025) -> 4.66 --------
In [8]:
viscosity_train = pd.read_csv("data/viscosity_train.csv", sep=";", decimal=",")
viscosity_test = pd.read_csv("data/viscosity_test.csv", sep=";", decimal=",")
display(viscosity_train.head(3))
display(viscosity_test.head(3))
| T | Al2O3 | TiO2 | Viscosity | |
|---|---|---|---|---|
| 0 | 20 | 0.0 | 0.0 | 3.707 |
| 1 | 25 | 0.0 | 0.0 | 3.180 |
| 2 | 35 | 0.0 | 0.0 | 2.361 |
| T | Al2O3 | TiO2 | Viscosity | |
|---|---|---|---|---|
| 0 | 30 | 0.0 | 0.0 | 2.716 |
| 1 | 40 | 0.0 | 0.0 | 2.073 |
| 2 | 60 | 0.0 | 0.0 | 1.329 |
In [9]:
from src.rules import simplify_and_group_rules
clustered_rules = simplify_and_group_rules(
viscosity_train, rules, clusters_num, kmeans.labels_
)
clustered_rules
Out[9]:
[[if (T = 70) and (TiO2 = 0.0) and (Al2O3 = 0.0) -> 1.194, if (T = 48.75) and (TiO2 = 0.0) and (Al2O3 = 0.0) -> 1.548, if (T = 42.5) and (TiO2 = 0.0) and (Al2O3 = 0.0) -> 1.629, if (T = 40.0) and (TiO2 = 0.0) and (Al2O3 = 0.0) -> 2.097, if (T = 27.5) and (TiO2 = 0.0) and (Al2O3 = 0.0) -> 3.18, if (T = 20) and (TiO2 = 0.0) and (Al2O3 = 0.0) -> 3.707], [if (T = 70) and (TiO2 = 0.0) and (Al2O3 = 0.3) -> 1.728, if (T = 48.75) and (TiO2 = 0.0) and (Al2O3 = 0.3) -> 2.293, if (T = 40.0) and (TiO2 = 0.0) and (Al2O3 = 0.3) -> 3.842, if (T = 27.5) and (TiO2 = 0.0) and (Al2O3 = 0.3) -> 5.162, if (T = 20) and (TiO2 = 0.0) and (Al2O3 = 0.3) -> 6.67], [if (T = 70) and (TiO2 = 0.3) -> 1.91, if (T = 48.75) and (TiO2 = 0.3) -> 2.688, if (T = 40.0) and (TiO2 = 0.3) -> 3.958, if (T = 27.5) and (TiO2 = 0.3) -> 5.405, if (T = 20) and (TiO2 = 0.3) -> 7.132], [if (T = 70) and (TiO2 = 0.1) and (Al2O3 = 0.0) -> 1.289, if (T = 48.75) and (TiO2 = 0.1) and (Al2O3 = 0.0) -> 1.662, if (T = 42.5) and (TiO2 = 0.1) and (Al2O3 = 0.0) -> 2.109, if (T = 40.0) and (TiO2 = 0.1) and (Al2O3 = 0.0) -> 2.726, if (T = 27.5) and (TiO2 = 0.1) and (Al2O3 = 0.0) -> 4.236, if (T = 20) and (TiO2 = 0.1) and (Al2O3 = 0.0) -> 4.885], [if (T = 42.5) and (TiO2 = 0.0) and (Al2O3 = 0.1) -> 1.897, if (T = 40.0) and (TiO2 = 0.0) and (Al2O3 = 0.1) -> 2.682, if (T = 27.5) and (TiO2 = 0.0) and (Al2O3 = 0.1) -> 3.38, if (T = 20) and (TiO2 = 0.0) and (Al2O3 = 0.1) -> 4.66]]
In [10]:
import numpy as np
from skfuzzy import control as ctrl
import skfuzzy as fuzz
temp = ctrl.Antecedent(viscosity_train["T"].sort_values().unique(), "temp")
al = ctrl.Antecedent(np.arange(0, 0.3, 0.005), "al")
ti = ctrl.Antecedent(np.arange(0, 0.3, 0.005), "ti")
viscosity = ctrl.Consequent(np.arange(1.18, 3.71, 0.00001), "viscosity")
temp.automf(5, variable_type="quant")
temp.view()
al.automf(3, variable_type="quant")
al.view()
ti.automf(3, variable_type="quant")
ti.view()
viscosity.automf(5, variable_type="quant")
viscosity.view()
c:\Users\user\Projects\python\fuzzy\.venv\Lib\site-packages\skfuzzy\control\fuzzyvariable.py:125: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown fig.show()
In [11]:
from src.rules import get_fuzzy_rules
fuzzy_variables = {"Al2O3": al, "TiO2": ti, "T": temp, "consequent": viscosity}
fuzzy_rules = get_fuzzy_rules(clustered_rules, fuzzy_variables)
fuzzy_cntrl = ctrl.ControlSystem(fuzzy_rules)
sim = ctrl.ControlSystemSimulation(fuzzy_cntrl, lenient=False)
display(len(fuzzy_rules))
fuzzy_rules
19
Out[11]:
[IF (temp[higher] AND ti[low]) AND al[low] THEN viscosity[lower] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[average] AND ti[low]) AND al[low] THEN viscosity[low] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[low] AND ti[low]) AND al[low] THEN viscosity[high] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[lower] AND ti[low]) AND al[low] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[higher] AND ti[low]) AND al[high] THEN viscosity[low] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[average] AND ti[low]) AND al[high] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[low] AND ti[low]) AND al[high] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[lower] AND ti[low]) AND al[high] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF temp[higher] AND ti[high] THEN viscosity[low] AND aggregation function : fmin OR aggregation function : fmax, IF temp[average] AND ti[high] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF temp[low] AND ti[high] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF temp[lower] AND ti[high] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[higher] AND ti[average]) AND al[low] THEN viscosity[lower] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[average] AND ti[average]) AND al[low] THEN viscosity[average] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[low] AND ti[average]) AND al[low] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[lower] AND ti[average]) AND al[low] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[average] AND ti[low]) AND al[average] THEN viscosity[average] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[low] AND ti[low]) AND al[average] THEN viscosity[high] AND aggregation function : fmin OR aggregation function : fmax, IF (temp[lower] AND ti[low]) AND al[average] THEN viscosity[higher] AND aggregation function : fmin OR aggregation function : fmax]
In [12]:
sim.input["temp"] = 20
sim.input["al"] = 0.0
sim.input["ti"] = 0.0
sim.compute()
sim.print_state()
display(sim.output["viscosity"])
=============
Antecedents
=============
Antecedent: temp = 20
- lower : 1.0
- low : 0.0
- average : 0.0
- high : 0.0
- higher : 0.0
Antecedent: ti = 0.0
- low : 1.0
- average : 0.0
- high : 0.0
Antecedent: al = 0.0
- low : 1.0
- average : 0.0
- high : 0.0
=======
Rules
=======
RULE #0:
IF (temp[higher] AND ti[low]) AND al[low] THEN viscosity[lower]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[higher] : 0.0
- ti[low] : 1.0
- al[low] : 1.0
(temp[higher] AND ti[low]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[lower] : 0.0
RULE #1:
IF (temp[average] AND ti[low]) AND al[low] THEN viscosity[low]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[average] : 0.0
- ti[low] : 1.0
- al[low] : 1.0
(temp[average] AND ti[low]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[low] : 0.0
RULE #2:
IF (temp[low] AND ti[low]) AND al[low] THEN viscosity[high]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[low] : 0.0
- ti[low] : 1.0
- al[low] : 1.0
(temp[low] AND ti[low]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[high] : 0.0
RULE #3:
IF (temp[lower] AND ti[low]) AND al[low] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[lower] : 1.0
- ti[low] : 1.0
- al[low] : 1.0
(temp[lower] AND ti[low]) AND al[low] = 1.0
Activation (THEN-clause):
viscosity[higher] : 1.0
RULE #4:
IF (temp[higher] AND ti[low]) AND al[high] THEN viscosity[low]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[higher] : 0.0
- ti[low] : 1.0
- al[high] : 0.0
(temp[higher] AND ti[low]) AND al[high] = 0.0
Activation (THEN-clause):
viscosity[low] : 0.0
RULE #5:
IF (temp[average] AND ti[low]) AND al[high] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[average] : 0.0
- ti[low] : 1.0
- al[high] : 0.0
(temp[average] AND ti[low]) AND al[high] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #6:
IF (temp[low] AND ti[low]) AND al[high] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[low] : 0.0
- ti[low] : 1.0
- al[high] : 0.0
(temp[low] AND ti[low]) AND al[high] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #7:
IF (temp[lower] AND ti[low]) AND al[high] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[lower] : 1.0
- ti[low] : 1.0
- al[high] : 0.0
(temp[lower] AND ti[low]) AND al[high] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #8:
IF temp[higher] AND ti[high] THEN viscosity[low]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[higher] : 0.0
- ti[high] : 0.0
temp[higher] AND ti[high] = 0.0
Activation (THEN-clause):
viscosity[low] : 0.0
RULE #9:
IF temp[average] AND ti[high] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[average] : 0.0
- ti[high] : 0.0
temp[average] AND ti[high] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #10:
IF temp[low] AND ti[high] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[low] : 0.0
- ti[high] : 0.0
temp[low] AND ti[high] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #11:
IF temp[lower] AND ti[high] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[lower] : 1.0
- ti[high] : 0.0
temp[lower] AND ti[high] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #12:
IF (temp[higher] AND ti[average]) AND al[low] THEN viscosity[lower]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[higher] : 0.0
- ti[average] : 0.0
- al[low] : 1.0
(temp[higher] AND ti[average]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[lower] : 0.0
RULE #13:
IF (temp[average] AND ti[average]) AND al[low] THEN viscosity[average]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[average] : 0.0
- ti[average] : 0.0
- al[low] : 1.0
(temp[average] AND ti[average]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[average] : 0.0
RULE #14:
IF (temp[low] AND ti[average]) AND al[low] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[low] : 0.0
- ti[average] : 0.0
- al[low] : 1.0
(temp[low] AND ti[average]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #15:
IF (temp[lower] AND ti[average]) AND al[low] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[lower] : 1.0
- ti[average] : 0.0
- al[low] : 1.0
(temp[lower] AND ti[average]) AND al[low] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
RULE #16:
IF (temp[average] AND ti[low]) AND al[average] THEN viscosity[average]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[average] : 0.0
- ti[low] : 1.0
- al[average] : 0.0
(temp[average] AND ti[low]) AND al[average] = 0.0
Activation (THEN-clause):
viscosity[average] : 0.0
RULE #17:
IF (temp[low] AND ti[low]) AND al[average] THEN viscosity[high]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[low] : 0.0
- ti[low] : 1.0
- al[average] : 0.0
(temp[low] AND ti[low]) AND al[average] = 0.0
Activation (THEN-clause):
viscosity[high] : 0.0
RULE #18:
IF (temp[lower] AND ti[low]) AND al[average] THEN viscosity[higher]
AND aggregation function : fmin
OR aggregation function : fmax
Aggregation (IF-clause):
- temp[lower] : 1.0
- ti[low] : 1.0
- al[average] : 0.0
(temp[lower] AND ti[low]) AND al[average] = 0.0
Activation (THEN-clause):
viscosity[higher] : 0.0
==============================
Intermediaries and Conquests
==============================
Consequent: viscosity = 3.499157499995422
lower:
Accumulate using accumulation_max : 0.0
low:
Accumulate using accumulation_max : 0.0
average:
Accumulate using accumulation_max : 0.0
high:
Accumulate using accumulation_max : 0.0
higher:
Accumulate using accumulation_max : 1.0
np.float64(3.499157499995422)
In [13]:
def fuzzy_pred(row):
sim.input["temp"] = row["T"]
sim.input["al"] = row["Al2O3"]
sim.input["ti"] = row["TiO2"]
sim.compute()
return sim.output["viscosity"]
result_train = viscosity_train.copy()
result_train["ViscosityPred"] = result_train.apply(fuzzy_pred, axis=1)
result_train.head(15)
Out[13]:
| T | Al2O3 | TiO2 | Viscosity | ViscosityPred | |
|---|---|---|---|---|---|
| 0 | 20 | 0.00 | 0.0 | 3.707 | 3.499157 |
| 1 | 25 | 0.00 | 0.0 | 3.180 | 3.188565 |
| 2 | 35 | 0.00 | 0.0 | 2.361 | 2.732494 |
| 3 | 45 | 0.00 | 0.0 | 1.832 | 1.812498 |
| 4 | 50 | 0.00 | 0.0 | 1.629 | 1.812498 |
| 5 | 55 | 0.00 | 0.0 | 1.465 | 1.812498 |
| 6 | 70 | 0.00 | 0.0 | 1.194 | 1.390833 |
| 7 | 20 | 0.05 | 0.0 | 4.660 | 3.481064 |
| 8 | 30 | 0.05 | 0.0 | 3.380 | 3.090537 |
| 9 | 35 | 0.05 | 0.0 | 2.874 | 2.703435 |
| 10 | 40 | 0.05 | 0.0 | 2.489 | 2.365680 |
| 11 | 50 | 0.05 | 0.0 | 1.897 | 2.054459 |
| 12 | 55 | 0.05 | 0.0 | 1.709 | 2.128746 |
| 13 | 60 | 0.05 | 0.0 | 1.470 | 1.465795 |
| 14 | 20 | 0.30 | 0.0 | 6.670 | 3.499157 |
In [14]:
result_test = viscosity_test.copy()
result_test["ViscosityPred"] = result_test.apply(fuzzy_pred, axis=1)
result_test
Out[14]:
| T | Al2O3 | TiO2 | Viscosity | ViscosityPred | |
|---|---|---|---|---|---|
| 0 | 30 | 0.00 | 0.00 | 2.716 | 3.089540 |
| 1 | 40 | 0.00 | 0.00 | 2.073 | 2.359522 |
| 2 | 60 | 0.00 | 0.00 | 1.329 | 1.465795 |
| 3 | 65 | 0.00 | 0.00 | 1.211 | 1.414928 |
| 4 | 25 | 0.05 | 0.00 | 4.120 | 3.188565 |
| 5 | 45 | 0.05 | 0.00 | 2.217 | 2.045546 |
| 6 | 65 | 0.05 | 0.00 | 1.315 | 1.414928 |
| 7 | 70 | 0.05 | 0.00 | 1.105 | 1.408926 |
| 8 | 45 | 0.30 | 0.00 | 3.111 | 3.499157 |
| 9 | 50 | 0.30 | 0.00 | 2.735 | 3.475062 |
| 10 | 65 | 0.30 | 0.00 | 1.936 | 1.812498 |
| 11 | 30 | 0.00 | 0.05 | 3.587 | 3.111691 |
| 12 | 55 | 0.00 | 0.05 | 1.953 | 2.128746 |
| 13 | 65 | 0.00 | 0.05 | 1.443 | 1.414928 |
| 14 | 40 | 0.00 | 0.30 | 3.990 | 3.475062 |
| 15 | 50 | 0.00 | 0.30 | 3.189 | 3.475062 |
| 16 | 65 | 0.00 | 0.30 | 2.287 | 1.812498 |
In [15]:
import math
from sklearn import metrics
rmetrics = {}
rmetrics["RMSE_train"] = math.sqrt(
metrics.mean_squared_error(result_train["Viscosity"], result_train["ViscosityPred"])
)
rmetrics["RMSE_test"] = math.sqrt(
metrics.mean_squared_error(result_test["Viscosity"], result_test["ViscosityPred"])
)
rmetrics["RMAE_test"] = math.sqrt(
metrics.mean_absolute_error(result_test["Viscosity"], result_test["ViscosityPred"])
)
rmetrics["R2_test"] = metrics.r2_score(
result_test["Viscosity"], result_test["ViscosityPred"]
)
rmetrics
Out[15]:
{'RMSE_train': 1.0977710360150494,
'RMSE_test': 0.4076186194536602,
'RMAE_test': 0.5797504263400755,
'R2_test': 0.813200460937507}