CHAPTER - 1

INTRODUCTION

    1.1     OVERVIEW OF THE PROJECT

Parkinson's disease is the world’s second-ranking common brain disease which affects brainpower and results in unintended body movements. Early signs of Parkinson's disease are very much difficult to notice. The main cause of Parkinson’s disease is the deficiency of the hormone dopamine that is present in the brain. It is produced by brain cells. it acts as a message-bearer between brain neurons and other body parts. The loss of dopamine affects the transmission of messages between the brain and other body parts.

Main reasons for causing Parkinson’s Disease are:

Age: Parkinson's disease is a rare occurrence among young adults. It usually starts in middle or late life, and the risk gets higher as you get older. The disease typically strikes people at 60 or older. Making family planning decisions may be aided by genetic counselling if a young person is diagnosed with Parkinson's disease. Also distinct from those of an older individual with Parkinson's disease and requiring special consideration are work, social circumstances, and drug side effects.

 Heredity: The likelihood that you'll develop Parkinson's disease increases if you have close family members who have the condition. Unless you have a large number of family members who suffer from Parkinson's disease, your risks are still minimal. and adverse medication reactions.

Head Injuries: Damaging of brain nerves due to head injuries because off several accidents is one of the main reasons for Parkinson’s disease

The presence of Lewy bodies: Microscopical indicators of Parkinson's disease include clumps of particular chemicals within brain cells. Lewy bodies are what they are, and scientists think they offer a crucial insight into what causes Parkinson's disease.

Formation of tremors is usual in human beings. But some tremors are unnoticeable in early stages but becoming dangerous gradually. PD patients are attacked by one of the tremors, which causes to cumulative sickness and affects the nerves on time being. The tremors, which lead to PD, develop barely in one hand of the person, which may be inconspicuous at early stages. But they will be slowing the movements of limbs such as legs, hands, thumbs and develops stiffness in them. The facial expressions of the patient of PD will be downing gradually at early stages and they are not able to swing their arms while walking. Speech become mumble or slurred.

The symptoms of this disease become terrible over time. Despite the fact that there the PD can’t be cured but medication helps in reducing the effect of symptoms significantly, if the disease identified in an early stages. Sometimes, doctors may even advise surgery to regulate certain regions of the brain. Person affected by this disease May slowing lose the ability to give facial expressions as a normal person and may not be able to wave their hands while moving. They will speek indistinctly. The condition of the person may become dreadful over a period of time. In spite of that this sickness is incurable but early detection the disease helps in reducing the effect of this sickness. In some cases surgeries are also options to regulate some parts of brain.

At first limb experience the tremor, which is rhythmic shaking, typically the hand or fingers. That could wiggle your thumb and forefinger. Even when diseased persons are at ideal state, the hand may shake. As the patient work, the shaking might be less. PD may slow your movements over time, which makes challenging and long delayed. Changes in the overall health can also impact PD symptoms.

Some of the health concerns can be linked to an increase in symptoms or worsening of the disease like infections and medication changes. Eventually they are not able to take longer steps while walking. It may become difficult to get up from a chair. Firmness in muscles may occur in the body. This will cause lot of pain and the person will not be able to do the work or move as usual. This disease also affects the posture of the body, slowly lose the balance and fall down, might not be able to perform insentient movements like blinking, specking, smiling or swinging arms, specking quickly and eventually speech may become soft and slurry.

Symptoms will differ from individual. Sometimes these symptoms are barely noticeable in the starting stages of this disease. Even though symptoms get to affect the legs on either side of the body, they usually start affecting some position of the body and continue to become more complex there. The most common signs of this disease are tremors, slowed movements, inflexible muscles, weak posture and balance of the body, unable to perform unconscious movements, face difficulty in writing, irregularities in speech .

Fig 1.1 Symptoms of Parkinson’s Disease

The effect of PD will be different from one person to another. The symptoms of this disease are unnoticeable in the beginning. Tremor affects the lower limbs of one side the body and affects the posture of the body and become more vicious over time. The usual symptoms of this disease are reducing the speed of gestures, stubborn muscles, and frail body position and may not be able to do unconscious gestures and may become hard in writing and speaking. The posture of the body should be the main consideration when predicting the disease [5].

Parkinson's has four main symptoms:

·       Tremor in hands, arms, legs, jaw, or head.

·       Muscle stiffness, where muscle remains contracted for a long time.

·       Slowness of movement.

·       Impaired balance and coordination, sometimes leading to falls.

There are 5 stages of Parkinson's disease

Stage 1: Does not affect the daily routines of a person

Stage 2: Starts affecting the routine activities.

Stage 3: Takes more time to complete the routine works.

Stage 4: Difficult to do their own activities.

Stage 5: Confined to bed and the person will die in a short period of time. The most common problems caused by Parkinson's disease are

Tremor:

The hand or fingers are frequently the first part of the body to tremble or shake rhythmically. Your thumb and forefinger can be moved. This is referred to as a "pill- rolling tremor". Your hand could tremble even when it is at rest. The shaking could reduce while you working.

Rigid body movements:

Become very hard to move body parts.

Change in voice:

We can observe changes in the voice the means there will be some deflection in voice and hard speaking in advanced stages.

Development of quivering is common with increasing age but some tremors are inconspicuous in the early stage but may lead to vicious consequences. Tremors developed in the patients affected by PD cause aggregative sickness in the person affecting the nervous system. This kind of tremors will affect the movement of the limbs such as hands, legs and thumbs and develop rigidity in the PD person [6]. In most cases, when PD is diagnosed, it has already developed to an advanced stage, there has been significant neuron loss and damage, at that stage any treatment will be may not be able to stop the disease from spreading further or lead to neuroprotection. The main goal of this study is to identify the characteristics of limb movements of the person affected by this disease, and by using machine learning the status of the disease can be identified with earlydetection, the person affected with disease can treated so that the person can live longer [6].

Since there is no unassailable test for PD, the prognosis of the disease must now only be made using clinical and observational criteria. Many of the symptoms of PD are difficult to identify and are alsoseen in other disorders. To assess the severity of the condition, a doctor will use the UPDRS, a measure based on a score produced from the neurological evaluation. However, because it is a qualitative perception, the scale lacks objectivity, reproducibility and sensitivity [7].

The most prevalent pre-diagnostic symptom of PD istremor, which 41% of patients report to their doctors in the preceding two to diagnosis, compared to lessthan 1% of controls. At five and ten years prior to diagnosis, the incidence of tremor are already high [7].Machine learning algorithms have shown promise in detecting PD based on various biomarkers, including spiral drawings and voice recordings[8].

Spiral drawings are a common biomarker for PD, as the tremors associated with the disease can cause irregularities in the drawing pattern. Machine learning algorithms can analyze these patterns and identify the presence of PD with high accuracy [8].Similarly, voice recordings can also be used to detect PD, as the disease can affect the vocal cords and cause changes in speech patterns. By analyzing features of the voice, such as pitch and tremors, machine learning algorithms can identify the presence of PD [9].

XGBoost is a popular machine learning algorithm for classification of tasks.. XGBoost is a gradient boosting algorithm that also uses an ensemble of decision trees but focuses on optimizing the gradient descent algorithm to minimize the loss function [9]. The DenseNet (densely connected convolutional network) is recognized for having convolutional neural network architecture that is state-of-art, when validated for classification using the popular ImageNet dataset. Huang et al. validated the technique of using direct connections in a feed-forward manner from each layer to every other layer. Every layer in the model architecture takes the target input and concatenation of the preceding layers’ feature maps. A DenseNet is a type of convolutional neural network that utilises dense connections between layers, through Dense Blocks, where we connect all layers (with matching feature-map sizes) directly with each other. It performs non-linear operations such as batch normalization, ReLU, and convolution or pooling. If the size of the feature maps changes, the concatenation procedure is unsuccessful. [9].

1.2     SCOPE

PD is a neurodegenerative disorder that affects the motor system, leading to symptoms such as tremors, rigidity, and bradykinesia. Early diagnosis and treatment of PD are crucial for slowing down the progression of the disease and improving the quality of life of patients [8].

The scope behind using Densenet and XGBoost for PD detection is to create a model that can accurately diagnose the disease based on patient data. This data may include information such as age, gender, medical history, and symptoms. Spiral drawings are a common biomarker for PD, as the tremors associated with the disease can cause irregularities in the drawing pattern. Machine learning algorithms can analyze these patterns and identify the presence of PD with high accuracy [8]. By training the machine learning model on a large dataset of patient information, it can learn to identify patterns and correlations between these variables and the likelihood of PD [9].

Once the model is trained, it can be used to analyze new patient data and provide a diagnosis. Thiscan help doctors to make a more accurate diagnosis, potentially at an earlier stage of the disease, which can lead to better treatment outcomes [9].

The use of machine learning algorithms on spiral drawings and voice recordings has several advantages over traditional diagnostic methods such as neurological exams and brain imaging. These methods can be time-consuming, expensive, and invasive. Using machine learning algorithms can lead to faster and more accurate diagnosis of PD, which can help patients receive timely treatment [10].

1.3     INTRODUCTION TO MACHINE LEARNING

Machine Learning is the sub-area of AI, whereby the term refers to the power of IT systems to independently find solutions to problems by recognizing patterns in databases. Machine Learning is an AI technique that teaches computers to learn from experience. Machine learning algorithms use computational methods to “learn” information directly from data without relying on a predetermined equation as a model. In other words: Machine Learning enables IT systems to acknowledge patterns in the idea of existing algorithms and data sets and to develop adequate solution concepts. Therefore, in Machine Learning, artificial knowledge is generated on the idea of experience. While artificial intelligence and machine

learning are often used interchangeably, they are two different concepts. AI is the broader concept – machines making decisions, learning new skills, and solving problems in a similar way to humans – whereas machine learning is a subset of AI that enables intelligent systems to autonomously learn new things from data. Instead of programming machine learning algorithms to perform tasks, you can feed them examples of labeled data (known as training data), which helps them make calculations, process data, and identify patterns automatically.

In order to enable the software to independently generate solutions, the prior action of 5 people is important. For example, the required algorithms and data must be fed into the systems in advance and the respective analysis rules for the recognition of patterns in the data stock must be defined.

Once these two steps have been completed, the system can perform the following tasks by Machine Learning [10]:

Ø Finding, extracting and summarizing relevant data

Ø Making predictions based on the analysis data

Ø Calculating probabilities for specific results

Basically, algorithms play a crucial role in Machine Learning: On the one hand, they're liable for recognizing patterns and on the opposite hand, they will generate solutions. Algorithms can be divided into different categories [11].

1.3.1  Supervised Learning

Techniques of Supervised Machine Learning algorithms include linear and logistic regression, multi-class classification, Decision Tree, and Support Vector Machine. Supervised Learning problems area kind of machine learning technique often further grouped into Regression and Classification problems. The difference between these two is that the dependent attribute is numerical for regression and categorical for classification [11].

Ø Regression : Linear regression could also be a linear model, e.g. a model that assumes a linear relationship between the input variables (x) and thus the only output variable (y). More specifically, that y is usually calculated from a linear combination of the input variables (x). Whenthere's one input variable (x), the tactic is mentioned as simple

linear regression. When there are multiple input variables, literature from statistics often refers to the tactic as multiple linear regression [11].

Ø Classification : Classification could also be a process of categorizing a given set of data into classes, It is often performed on both structured or unstructured data. the tactic starts withpredicting the category of given data points. The classes are often mentioned as target, label, or categories. In short, classification either predicts categorical class labels or classification data supported the training set and thus the values(class labels) in classifying attributes and uses it in classifying new data. There is a variety of classification models. Classification models include Logistic Regression, Decision Tree, Random Forest, Gradient Boosted Tree, One-vs.-One, and Naïve Bayes [11].

1.3.2  Unsupervised Learning

There In unsupervised learning, AI learns without predefined target values and without rewards. It's mainly used for learning segmentation (clustering). The machine tries to structure and type the info entered consistent with certain characteristics. Unsupervised Learning is that the training of Machines using information that's neither classified nor labeled and allowing the algorithm to act there on information without guidance.

Unsupervised Learning is accessed into two categories of algorithms [11]:

Ø Clustering : A clustering problem is where you would like to get the inherent grouping in the data such as grouping customers by purchasing behavior.

Ø Association : An Association rule learning problem is where it would wish to get rules that describe large portions of your data such as folks that buy X also tend to shop for Y.

1.3.3  Classification Metrics

Classification metric is a number that measures the performance that your machine learning model when it comes to assigning observations to certain classes. Binary classification is a particular situation where you just have to classes both positive and negative. Classification is a supervised learning task in which it will try to predict the class or label of a data point based on some feature values.Depending on the number of

classes target variable includes, it can a binary or multi-class classification. Evaluating a machine learning model is just as important as building it.

Classification measurements are the type of measurements which establish the a priori ill-defined composition and essence of the designations (classes) of a categorical measurement scale and the result is the assignment of a single object to one of the chosen classes. The most straight forward way to measure a classifier's performance is using the Accuracy metric. Here, we compare the actual and predicted class of each data point, and each match counts for one correct prediction. Accuracy is then given as the number of correct predictions divided by the total number of predictions. Evaluation metrics are tied to machine learning tasks. Using different metrics for performance evaluation, we should be able to improve our model’s overall predictive power before we roll it out for production on unseen data. Without doing a proper evaluation of the Machine Learning model by using different evaluation metrics, and only depending on accuracy, can lead to a problem when the respective model is deployed on unseen data and may end in poor predictions. classification metric is a number that measures the performance that your machine learning model when it comes to assigning observations to certain classes. Binary classification is a particular situation where you just have to classes: positive and negative. The optimal choice of metric usually depends on the characteristic of the data and the given task. It will also explain when a particularmetric might be a better fit than the others [12].

1. True Positives : It is the case where we predicted Yes and the output is also Yes.

2. True Negatives : It is the case where we predicted No and the output is also No.

3.    False Positives : It is the case where we predicted Yes but, it is actually No.

4.    False Negatives : It is the case where we predicted No but, it is actually Yes.

The various machine learning classification metrics for different algorithms are given below:

Accuracy: One parameter for evaluating classification models is accuracy. Accuracy is defined as the percentage of correct predictions for the test data. It can be calculated easily by dividing the number of correct predictions by the number of total predictions. The percentage of predictions that the model correctly predicted is known as accuracy. The following is the actual definition of accuracy [12].

For binary classification, accuracy can also be calculated in terms of positives and negatives as follows:

Sensitivity: It is a measure of the proportion of actual positive cases that got predicted as positive (or true positive). Sensitivity is used to evaluate model performance because it allows us to see how many positive instances the model was able to correctly identify. This implies that there will be another proportion of actual positive cases, which would get predicted incorrectly as negative (or false negative) [12].

Specificity: It is defined as the proportion of actual negatives, which got predicted as the negative (or true negative). This implies that there will be another proportion of actual negative, which got predicted as positive and could be termed as false positives. This proportion could also be called a false positive rate [12].

Precision: It is the Ratio of true positives to total predicted positives.

Recall: It is calculated as the proportion of Positive samples that have been accurately identifiedas Positive to all Positive samples. The recall evaluates how well the model can identify positive samples. The more positive samples that are identified, the greater the recall [12].

F1-Score: It is a harmonic mean between recall and precision. Its range is [0,1 This metric typically indicates how precise (how many times it correctly classifies) and how robust the system is. It is used to measure the test’s accuracy [12].