Parkinson’s Neuro-disease Exploratory Data Analysis and prediction with machine Learning
Parkinson's disease (PD), or simply Parkinson's is a long-term degenerative disorder of the central nervous system that mainly affects the motor system. The symptoms usually emerge slowly and, as the disease worsens, non-motor symptoms become more common.[1][4] The most obvious early symptoms are tremor, rigidity, slowness of movement, and difficulty with walking,[1] but cognitive and behavioral problems may also occur. Parkinson's disease dementia becomes common in the advanced stages of the disease. Depression and anxiety are also common, occurring in more than a third of people with PD.[2] Other symptoms include sensory, sleep, and emotional problems. The main motor symptoms are collectively called "parkinsonism", or a "parkinsonian syndrome". [1] In this lexture, we are going to perform and exploratory data analysis and code a machine learning model to predict if a person have this disease. 1- Dataset The dataset used is a dataset of the web site machine Learning repository[2] which contains a high quality of machine Learning dataset. The data is composed of the following features: - name - ASCII subject name and recording number - MDVP:Fo(Hz) - Average vocal fundamental frequency - MDVP:Fhi(Hz) - Maximum vocal fundamental frequency - MDVP:Flo(Hz) - Minimum vocal fundamental frequency -MDVP:Jitter(%),MDVP:Jitter(Abs),MDVP:RAP,MDVP:PPQ,Jitter:DDP – Several - measures of variation in fundamental frequency - MDVP:Shimmer,MDVP:Shimmer(dB),Shimmer:APQ3,Shimmer:APQ5,MDVP:APQ,Shimmer:DDA - Several measures of variation in amplitude - NHR,HNR - Two measures of ratio of noise to tonal components in the voice - status - Health status of the subject (one) - Parkinson's, (zero) – healthy - RPDE,D2 - Two nonlinear dynamical complexity measures DFA - Signal fractal scaling exponent spread1,spread2,PPE - Three nonlinear measures of fundamental frequency variation Separation of the dataset For analysis we separed the data in 2 sets.
Analysing Gender.
The gender in the dataset is distributed as follows:
ax = sns.countplot(x='class',data=data, hue = 'gender')
ax.set_title('Frequency of Gender by Class')
1- violin plot of Formant frequency
Analysis of Formant Frequencies by gender
We have obtained the following results:
frequency_ill= illness[['gender','f1','f2','f3','f4']]
frequency_health = healthy[['gender','f1','f2','f3','f4']]
frequency_ill.head() frequency_ill['mean'] = frequency_ill[['f1','f2','f3','f4']].mean(axis=1) frequency_health['mean'] = frequency_health[['f1','f2','f3','f4']].mean(axis=1)
fig, axes = plt.subplots(1,2, sharex = True)
sns.swarmplot(x='gender', y='mean',data=frequency_ill, ax = axes[0])
sns.swarmplot(x='gender', y='mean',data=frequency_health, ax = axes[1])
axes[0].set_title("frequency of parkingson's positive")
axes[1].set_title("frequency of parkingson's negative") plt.tight_layout() fig, axes = plt.subplots(2,1, sharex = True)
sns.violinplot(x='gender', y='mean',data=frequency_ill, ax = axes[0])
sns.violinplot(x='gender', y='mean',data=frequency_health, ax = axes[1]) axes[0].set_title("frequency of parkingson's positive")
axes[1].set_title("frequency of parkingson's negative")
plt.tight_layout()
2- Boxplot
fig, axes = plt.subplots(1,2, figsize=(8,6))
sns.boxplot(x='gender', y='mean',data=frequency_ill, ax = axes[0])
sns.boxplot(x='gender', y='mean',data=frequency_health, ax = axes[1])
plt.ticks_layouts()
plt.show()
Analysis of intensity parameters
sns.distplot(illness['meanIntensity'])
sns.distplot(healthy['meanIntensity'])
table = list(np.arange(0,100,5))
percentiles_i = []
percentile_h = []
for i in table :
perc = np.round(np.percentile(illness['meanIntensity'], i),2)
per = np.round(np.percentile(healthy['meanIntensity'],i),2)
percentiles_i.append(perc)
percentile_h.append(per)
sns.distplot(percentiles_i,hist=False)
sns.distplot(percentile_h,hist=False)
plt.title('Intensity Distribution for positive and negative parkingson disease')
II-predicting Parkinson’s disease with machine Learning.
1- Importing packages
from sklearn.ensemble import RandomForestClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.feature_selection import VarianceThreshold
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.model_selection import train_test_split,GridSearchCV
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score,f1_score
from sklearn.svm import SVC
2- Prediction with logistic Regression
The data has been split with train_test_plit and standardized.
NB. Metrics used are f1_score and accuracy.
logistic = DecisionTreeClassifier(class_weight = 'balanced',random_state = 11) logistic.fit(X_train,y_train)
prediction = logistic.predict(X_test) accuracy_score(prediction, y_test).
The accuracy score with logistics regression is 0.86
f1_score(rf_prediction, y_test)
The f1_score with logistics regression is 0.87
3- Prediction with Random Forest Classifier
rf = RandomForestClassifier() rf.fit(X_train,y_train) rf_prediction = rf.predict(X_test) accuracy_score(rf_prediction, y_test)
Accuracy with random Forest is 0.92
f1_score(rf_prediction, y_test)
The f1_scoreis 0.95
4- Prediction with SVM.
sv = SVC(class_weight = 'balanced') sv.fit(X_train,y_train)
prediction = sv.predict(X_test) accuracy_score(prediction, y_test) f1_score(prediction,y_test)
With SVM we have obtained an accuracy of 0.86 and a f1_score of 0.92
5- Hyper parameters tuning
from scipy.stats import uniform
C = [1.0,1.5,2.0,2.5]
param_grid = dict(C=C)
lr = LogisticRegression(penalty='l2')
grid = GridSearchCV(estimator = lr, param_grid=param_grid,scoring = 'accuracy', cv = 4,n_jobs = -1)
grid.fit(X_train, y_train)
y_pred = grid.predict(X_test)
f1_score(prediction, y_test)
grids.score(X_test, y_test)
Accuracy: 0.81
F1_score:0.87
n_estimators = [10,100,200]
max_features = [4,5,8]
params_grid = dict(n_estimators = n_estimators, max_features = max_features)
rfc = RandomForestClassifier()
search = GridSearchCV(estimator = rfc, param_grid=params_grid, cv = 3, scoring = 'accuracy', n_jobs=-1)
search.fit(X_train,y_train)
search.score(X_test, y_test)
f1_score(pred, y_test) kernels
Accuracy:0.83
F1_score:0.88
['poly','rbf','sigmoid']
C = [0.1,10,100]
gamma = [1,0.1,0.01,0.001]
param_grid = dict(C=C, kernel = kernels, gamma = gamma)
svm = SVC()
grids = GridSearchCV(estimator = svm, param_grid = param_grid, scoring='accuracy', cv = 3, n_jobs = -1)
grids.fit(X_train,y_train)
prediction =grids.predict(X_test)
grids.score(X_test, y_test)
f1_score(prediction, y_test)
Accuracy: 0.95
F1_score:0.96.
After a grid search we conclude that the Best model is a support vector machine with and accuracy of 0.95 and a F1_score of 0.96.
6- Conclusion.
We have implemented a model which can predict the presence of Parkinson’s disease with a accuracy of 0.95 and with F1_score of 0.96.
Bibliography:
1. Wikipedia.com/Parkinson’s disease