Page 235 - Big Data Analytics for Intelligent Healthcare Management

228 CHAPTER 9 INTELLIGENCE-BASED HEALTH RECOMMENDATION SYSTEM

can be described as being low-risk, choices made in other sectors may have more intense ramifications
for the end user. In particular, in the healthcare sector, choices can be life-threatening as they can con-
cern the life and safety of patients. The recommendation system should not only support decision-
making and avert dangers and failures, but it should also monitor patients and dispense treatment
as necessary, keep track of vital signs, and communicate in real time via a centralized server. This in-
creases the suitability of health recommendation systems [12, 13].
The rest of the chapter is organized as follows: Section 9.2 presents an overview of the recommen-
dation system, its basic concepts, and different filtering techniques along with its evaluation.
Section 9.3 describes the health recommendation system framework, methods, and its evaluation.
Section 9.4 discusses the proposed intelligent health recommendation system along with experimental
results. Section 9.5 shows the advantages and disadvantages of the proposed HRS and Section 9.6 com-
prises the conclusion and future work.

9.2 BACKGROUND
The recommendation system is one type of decision-making system that decides which item can be
recommended to the user based on a similarity measure among items or users. The basic concepts
of the recommendation system follow the phases behind the development of systems and different fil-
tering techniques used in this decision system. This system helps healthcare organizations to improve
their own systems by recommending the right product to the user.

9.2.1 RECOMMENDATION SYSTEM AND ITS BASIC CONCEPTS
In the recommendation system, two main entities play the main role, i.e., products and end users. Users
provide their preferences about certain items and using the collected data, these preferences can be
found. The collected data are portrayed as a utility matrix that provides the value of each user-item
pair that represents the degree of preferences of that user for specific items. In this way, these are
mainly two types of recommendation systems: user-based and item-based. In the user-based recom-
mendation system, users provide their choices and ratings on items. We can recommend that item
to the user, which is not rated by that user with the help of user-based recommender engine, considering
similarity among the users. In the item-based recommender system, we use similarity between items
(not users) to make predictions from users. Data collection for a recommendation system is the first job
for prediction.

9.2.2 PHASES OF RECOMMENDATION SYSTEM
• Information collection phase
This phase gathers vital information about users and prepares user profiles based on the users’ attribute,
behaviors, or resources accessed. Without constructing a well-defined user profile, the recommenda-
tion engine cannot work properly. A recommendation system is based on inputs that are collected in
different ways, such as explicit feedback, implicit feedback, and hybrid feedback. Explicit feedback
takes input given by users according to their preferences for a product whereas implicit feedback