Based on the estimated facts, World Health Organization (WHO) had claimed that about 17.9 million men and women died from Cardio Vascular Disease (CVD) in year 2016, which is nearly 31% of the deaths worldwide [1]. Globally, CVD is the major cause of death in men and women who die annually from CVD [2]. It is also reported that one third of the deaths for people aged below 70 years occur by stroke and heart attacks [3]. Though there are leaping advances in medical field, it is still challenging to predict the heart disease in an earlier stage. Antenatal practices help in identifying the risk factors and modern technology is used to overcome the risk [4]. Still, the morbidity rate is very high due to the unpredicted and unexpected reasons. Currently, there are no proven and concrete interventions that reduce the risk of cardiovascular disease [5,6].

In cardiovascular disease, the data provided to the medical field is vast, multi-dimensional and critical, which makes it challenging and cumbersome to understand such extensive data. It is quite difficult to predict whether the patient has heart disease symptoms or not based on the patient's records [7,8]. Some of the typical symptoms for patients having CVD are chest pain, shortness of breath, radiating pain in the arms, left shoulders and elbows, discomfort while walking, high blood pressure, dizziness, nausea, fatigue, etc. while the risk factors causing CVD are cholesterol, hypertension, genetic or heredity, obesity, diabetes, dietary habits, aging, [9–11], etc. Life Style modifications with regular cardio friendly exercises followed by a heart-healthy diet, quitting smoking and alcohol intake, usage of tobacco, salt, sugar and fat intake can prevent CVD. Based on the analysis of worldwide data on CVD patients, it is noted that CVD affects even young men and women and therefore, requires prediction and prevention at the incipient stage itself.

The key objective of this research is to predict cardiovascular disease at an earlier stage. To envisage the detection of CVD at the incipient stage and intervene appropriately, Support Vector Machine (SVM) is used for classifying the patients depending on their symptoms and risk factors. By using these machine learning algorithms, it becomes easy to understand the nature and type of heart disease in all aspects.

The important aspect of this method is to train the machine to analyze a massive set of data with the known inputs and outputs, i.e., Supervised machine learning is applied [12]. The main objective of the work is to improve the accuracy of prediction of cardiovascular disease [13,14], which can help for preventive treatments and reduce the cases having CVD.

The rest of the paper deals with various sections viz.., the related works are discussed in Section 2, while in Section 3, the various stages of CVD are elaborated and the techniques involved for the prediction of cardiovascular disease are explained thoroughly. Section 4 highlights the Experimental results, Comparison of results using multiple techniques, Summarization of the methods and percentage of CVD based on the collected datasets. Finally, in Section 5, the conclusion of research and the thrust for more related research work is highlighted.

The clinical study was designed as a controlled clinical study, open-label [15]. To estimate the event-free rate, Kaplan-Meier has been used [16,17]. The logistic regression (LR) model reveals that a Coronary Artery Calcium (CAC) score of >=100 rather than a Cardio Ankle Vascular Index (CAVI) of >=9 had a higher predictive value for all cardiovascular events [18]. The absolute risk was calculated using charts and adjusted according to these notes. More than 20% chances of developing a cardiovascular event due to the prevalence of WHO/ISH of high CVD risk [19]. From the Receiver Operating Characteristic (ROC), a cross-sectional study was conducted. The risk of carotid arteriosclerosis was positively associated with CAVI [20].

The study includes neural networks, support vector machine, Decision Tree models, ID3 algorithm, and association rules [21]. To analyze the application of data mining algorithms, a study has been done [22]. Principal Component Analysis (PCA) is a method in which the large data set is reduced by PCA, wherein the original data is transformed into new dimensions. The bivariate model was used to find the relationship between two variables. In the future, the research can be extended by using different techniques like an outlier and link analysis used to associate rule mining on a large number of patients. Usage of various techniques and association of rule mining has been referred for many patients [23,24]. The methods used here are classification methods such as DT and SVM. The results present a better accuracy of 99.42% by using the decision tree, whereas the support vector machine gives an accuracy of 89% [25].

The KNN algorithm is based on the similarity measure and the distance has been evaluated [26]. To develop predictive models, machine learning techniques like SVM, DT and ANN are used. The main aim is to compare the three algorithms using the performance measure algorithms like sensitivity, accuracy, and specificity [27]. Nephritis of renal pelvis origin is a disease that affects the urinary system [28]. The machine learning algorithms used here are SVM, Naive Bayes, Logistic regression, J48, and one R. The highest accuracy is still unable to exceed the accuracy of 80% with the best performing algorithm [29]. The techniques used here are machine learning, such as logistic regression and classification trees [30].

Using ANN, the multi-layer feed-forward neural network with Levenberg-Marquardt learning algorithm was used on CVD and diabetes mellitus patients [31,32]. The proposed system's primary method was a neural network ensemble method.89.01% of classification accuracy has been taken from the Cleveland Heart Disease database [33]. Iris image was cropped from the region of interest, and the discrete wavelength transformation feature, statistical, texture is extracted from the area of interest [34].

In the proposed system, the main aim is to investigate the CAD system for breast cancer. The database used here was the Springer link (SL), inclusion and exclusion criteria [35]. Data mining methods, like feature extraction and selection, have been used. K means and SVM has been used for diagnosing tumors and extracting useful information [36]. Based on the classification technique, machine learning had been approached. For solving the classification of problems, the Genetically Optimized Neural Network (GONN) algorithm had been used [37]. In the proposed system, by using the data, the prediction of chronic disease has been done by using machine learning algorithms. MLT like KNN, SVM, decision tree and logistic regression has been used [38]. In the proposed research, the pre-processing techniques like removal of missing data, deletion of noisy data, if applicable, will fill the default values. The performance has been measured using methods like classification, sensitivity, accuracy and specificity analysis [39–41]. Advanced data mining techniques has been used to discover the relationship and hidden pattern. The dataset used here was the Iranian Center for Breast Cancer (ICBC) [42–44].

In the past several years, Cardio vascular disease (CVD) is one of the main contributing factors that cause innumerable fatalities worldwide. The medical condition that affects the heart and blood vessels is called CVD. Risk assessment and risk predictions have become essential in preventing CVD. Even though risk prediction tools are recommended, they are not adequately implemented in clinical practice. If the CVD patient is not properly assessed and intervened on time, there is every chance of resulting in fatality. To reduce the mortality rate and improve the heart health conditions, a new method is proposed. This proposed method analyzes the parameters (hypertension, diabetes, obesity, blood pressure, heredity), and also the typical and atypical symptoms associated with the heart patients. To identify and classify the stages of CVD affected patients, new classification methods are designed. The proposed classification methods have high accuracy in predicting CVD patients. It also has a fast-training process and therefore easy to indicate whether the patients are affected by CVD or not.

The various stages of CVD are shown in Fig. 1 constituting various stages as

In this, SVM classification algorithm are used to analyze the health data for cardiovascular disease. For each attribute in SVM, the optimal hyperplane has been predicted. The performance metrics such as Accuracy, Specificity and Sensitivity were used to prove the efficacy of this modern approach. Tab. 2 gives the baseline clinical characteristics of patients.

In this study the occurrence of CVD is more in humans because of genetic disorder when compared to other health issues. Therefore, medical practioner are in need of suitable technique or model to predict primarily. Nowadays, machine learning techniques are adapted in medical science for such decision-making situations. The proposed system designed under supervised learning using SVM plays efficiently in decision appealing and retrieving proficiency from an enormous amount of data thereby delivering better patient care and productivity. This system proves as an excellent model in addressing the classification problems and go-to method for any type of prediction problem. In this research paper, the classification techniques used for identifying cardiovascular in patients gives more prediction accuracy compare to other technical approaches. Future research is required to validate the algorithm for a huge population in taking many risk parameters into consideration.