Cloud Based Recommendation Services for Healthcare
More InformationShow full item record
With the inception of portable computing devices, enormous growth in the healthcare data over the Internet has been observed. Consequently, the Web based systems come across several challenges, such as storage, availability, reliability, and scalability. By employing the cloud computing to offer healthcare services helps in overcoming the aforementioned challenges. Besides the healthcare organizations, cloud computing services are also equally beneficial for general public in devising patient-centric or user-centric methodologies that involve users in managing health related activities. This dissertation proposes methodologies to: (a) make risk assessment about diseases and to identify health experts through social media using cloud based services, (b) recommend personalized health insurance plans, and (c) secure the personal health data in the cloud. The proposed disease risk assessment approach compares the profiles of enquiring users with the existing disease specific patient profiles and calculates the risk assessment score for that disease. The health expert consultation service permits users to consult with the health specialists that use Twitter by analyzing the tweets. The methodology employs Hyperlink-Induced Topic Search (HITS) based approach to distinguish between the doctors and non-doctors on the basis of tweets. For personalized health insurance plans identification, a recommendation framework to evaluate different health insurance plans from the cost and coverage perspectives is proposed. Multi-attribute Utility Theory (MAUT) is used to permit users evaluate health insurance plans using several criteria, for example premium, copay, deductibles, maximum out-of-pocket limit, and various other attributes. Moreover, a standardized representation of health insurance plans to overcome the heterogeneity issues is also presented. Furthermore, the dissertation presents a methodology to implement patient-centric access control over the patients’ health information shared in the cloud environment. This methodology ensures data confidentiality through the El-Gamal encryption and proxy re-encryption approaches. Moreover, the scheme permits the owners of health data to selectively grant access to users over the portions of health records based on the access level specified in the Access Control List (ACL) for different groups of users. Experimental results demonstrate the efficacy of the methodologies presented in the dissertation to offer patient/user-centric services and to overcome the scalability issues.