C. Kalaiselvi ¹, G. M. Nasira ²

Affiliations
1 Karpagam University, Coimbatore-641021, Tamilnad, IN
2 Department of Computer Science, Chikkanna Govt Arts College, Tiruppur-641602, Tamilnadu, IN

Abstract

Background: The heterogeneous, chronic diseases like heart diseases and cancer are commonly occur and increased nowadays in diabetic patients. Most of the people do not know the symptoms of these diseases and its chronic complications. Objective: The aim of this paper is to predict the diseases such as heart diseases and cancer in diabetic patients. The association between these diseases can be analyzed based on the factors that cause these diseases which include obesity, age, associated diabetic duration, and some other life style factors. Methods: This work consists of two stages. In the first stage, the attributes are identified and extracted using Particle Swarm Optimization (PSO) algorithm. In the second stage, ANFIS (Adaptive Neuro Fuzzy Inference System) with Adaptive Group based K-Nearest Neighbor (AGKNN) algorithm has been used to classify the data. Findings: The experimental results show a very good accuracy and signify the ANFIS with AGKNN along with feature subset selection using PSO. The performance is evaluated using performance metrics and proved this classifiers efficiency for the prediction of heart disease and cancer in diabetic patients. Application/ Improvement: This work demonstrates the diagnosis of diseases and its importance to predict it earlier. In future it can be implemented for other related diseases in medical data mining and healthcare.

Keywords

Classification, Data Mining, Disease Prediction, Feature Selection, Normalization

Full Text

   

S. Radhimeenakshi ¹, G. M. Nasira ²

Affiliations
1 Department of Computer Science, Tiruppur Kumaran College for Women, Bharathiar University, Tirupur - 641687, Tamil Nadu, IN
2 Department of Computer Applications, Chikkanna Government College, Tirupur - 641602, Tamil Nadu, IN

Abstract

Objective: The objective of this paper is to predict the risk level of Heart Disease by applying Probabilistic Neural Network trained with Particle Swarm Optimization in case of Remote Health Monitoring. Methods: In order to achieve the aim of the activity, we propose hybrid model of Particle Swarm Optimization (PSO) and Probabilistic Neural Network (PNN). PSO is a population based meta-heuristic Evolutionary Algorithm (EA) whose goal is to explore the search space in order to find near – optimal solutions for feature selection. The optimal features selected can be used for prediction system to develop a classification model using probabilistic Neural Network. Results: First, we quantify the clinical data set from the UCI machine learning repository and measured the complexity. There are 13 attributes are used such as the age which identifies the age of the person, chest pain type has 4 values, serum cholesterol level, blood sugar, resting ECG results, serum cholesterol level, amount of heart rate achieved, x old peak, number of major vessels colored by fluoroscopy, slope of the peak exercise ST segment, thal, sex, height, weight and additional factor smoking. It has been shown that the time complexity of hybridizing PSO and PNN obtained the promising results compared to other two algorithms such as regression tree and PSO optimization. We also proposed the data mining process to deal with complexity, missing values and high dimensionality followed by incorporating the data mining functionalities like characterization, discrimination, association, classification, prediction and evolution analysis. The experiment carried out in Java on stat log heart disease data set performs better in all noise conditions. Conclusion: The performance was evaluated in terms of time complexity, accuracy, sensitivity and specificity and it proved that the hybrid model of PSO and PNN outperformed the Regression tree and PSO.

Keywords

Expectation Maximization (EM), Heart Disease, Particle Swarm Optimization (PSO), Probabilistic Neural Network (PNN), Remote Heart Risk Monitoring System (RHRMS)

Full Text

   

R. Vidya ¹, G. M. Nasira ²

Affiliations
1 Department of Computer Science, MS University, Tirunelveli - 627012, Tamil Nadu, IN
2 Department of Computer Science, Chikkanna Government College, Tirupur – 641602, Tamil Nadu, IN

Abstract

Background/Objective: Cervical Cancer is one among the most vulnerable and highly affected diseases among women around the World. Normally, cells grow and divide to produce more cells only when the body needs them. This orderly process helps to keep dividing when new cells are not needed. These cells may form a mass of extra tissue called a growth of tumor. Tumors can be classified as Benign or Malignant. First Benign Tumors are not cancer. They can usually be removed, and in most cases, they do not show up. Most important, the cells in benign tumors do not spread to other parts of the body. Second, malignant tumors are cancer cells. These tumors can damage nearby tissues and organs. Malignant tumors are threat to life. In this research work, prediction of normal cervix or Cancer cervix is determined with the aid of Powerful Data Mining algorithms. Methods/Statistical Analysis: In this research work, prediction of normal cervix or Cancer cervix is determined with the aid of Powerful Data Mining algorithms. Data mining plays an indispensable role in prediction especially in medical field. Using this concept, Classification and Regression Tree algorithm, Random Forest Tree algorithm and RFT with K-means learning for prediction of normal cervix or cancer cervix is introduced. Collection of data from NCBI (National center for Bio-technology Information) in our work, we used the data set that contains 500 records and 61 variables (i.e. Biopsy numerical value with gene identifier). The output has been presented in the form of prediction tree format. As stated, we selected a sample of 100 records with 61 biopsy features. Based on this biopsy data, an awareness program is conducted and survey is followed up to identify the changes of women during this transition period. To collect data efficiently, a Personal Interview program was conducted among rural women in various places. Collaboration with JIPMER hospital people were checked up for the test of cervical cancer. The results obtained through biopsy test were put through statistical analysis and was given through MATLAB for algorithm testing. To ascertain the results obtained are segregated and delivered in various heads with 100 test data and 60 training data. Findings: Comparison of the performance of various algorithms was used under the techniques in terms of sensitivity, specificity and accuracy to determine the best predictor for the cervical cancer. At first, Regression tree algorithm methodology was used for prediction. The CART binary tree yields two results, either normal cervix or cancer cervix. A Splitting Criterion called GINI index is used to identify the diversity that exists in cervical data. RFT validated optimal accuracy, a new logic was applied i.e. "combinations of two algorithms" is used. It is also an ensemble supervised machine learning algorithm. The process of whitening is used as a pre-process in k-means clustering, to get the best prediction result. The result showed the 83.87% accuracy with CART TREE output. Random Forest Tree (RFT) is used to improve the prediction accuracy. With MATLAB Coding we achieved 93.54% of prediction accuracy. The K-Means algorithm is considered efficient for processing huge datasets and hence a high accuracy of 96.77% is achieved with RFT - K-MEAN LEARNING TREE output. The Randomization of Algorithm is presented in two ways: 1. Bagging for random bootstrap sampling and 2. Input attributes are selected at random for decision tree generation. This creates an unbiased estimate of generalization error as growing of tree into forest progressed and the derivation of time complexity of K-Means is achieved. Applications/Improvements: Cervical cancer diagnosis and prognosis are two medical applications which pose a great challenge to the researchers. The algorithms optimize a cost function defined on the Euclidean distance measure between the data points and means of cluster. Combination of RFT with K-means algorithm is the novelty of our research work, where we have achieved high accuracy result. Accurate prediction of occurrence of cervical cancer has been the most challenging and toughest task in medical data mining because of the non-availability of proper dataset. Many researchers have been done to develop different techniques that can solve problems and improve the prediction accuracy of cervical cancer through images. But in our research work, the prediction of cervical cancer is with Numerical Data. NCBI (National Center for Biotechnology Information) data set has been used. This research paper is a boon to create expert medical decision making systems and a solution for medical practitioners to construct an optimal prediction model for Cervical Cancer Prediction.

Keywords

Cervical Cancer, CART, Data Mining, Hereditary Pattern, K-Mean, RFT.

Full Text

   

V. Vallinayagi ¹, G. M. Nasira ²

Affiliations
1 M. S. University, Tirunelveli - 627 012, Tamil Nadu, IN
2 Chikkana Arts and Science College, Tirupur - 641 602, Tamil Nadu, IN

Abstract

In opportunistic routing protocol focus on the reliability of sending data packets from source to destination using different methods. Mobility forms an adequate challenge in networks. Generally routing protocols travel along trustworthy path and so problem may be created. A relay form of applicant set is selected for sending information. On each communication the hop which satisfies the selected criteria on receiving the packet will promote the data from source to destination. So we propose an efficient power, speed and link stability protocols that use the property of the location based routing and transmit wireless media. Selecting hops forms basic criteria for sending the data the proposed method will consume more energy and accommodating more hosts. We finally give common approach for accepting the nodes to select a path to destination.

Keywords

Adhoc Network, Magnet, Position based Routing Protocol, Reliability

Full Text

   

R. Umadevi ¹, G. M. Nasira ²

Affiliations
1 Department of Computer Science, Periyar University, Salem – 636011, Tamil Nadu, IN
2 Department of Computer Science, Chikkanna Government Arts College, Tirupur – 641602, Tamil Nadu, IN

Abstract

Now-a-days, several well-organized data hiding methods have been introduced and implemented using video steganography channel. Data hiding is to accomplish better data communication by hiding information into a video medium carrier to form an unrecognizable code stream. In the consideration of security aspect for communication process especially in video we enhanced this QoS factor security in this paper. Due to heavy packet loss, channel bit errors, video stego analysis have become a significant addition to the security aspects and are extensively developed as a complementary line to conventional security method. In this work, Fourier Transform With Security (FTWS) technique is applied for improve the security on video files in communication process. In this paper, a Secure Hash Polynomial Function to provide multi-layer security according to the modular additions and density functions is presented. Through the analysis of performance factor the security is important factor in our research. Finally, the simulation results evaluated by Video Quality Experts Group with parameter such as security with multilayered for 110 frames using MATLAB. It shows that the method FTWS enhance the security significantly compared with typical state-of-the-art methods.

Keywords

Fourier Transform, Hash Polynomial Function, Packet Loss, Security, Video Steganography

Full Text