Page 251 - Big Data Analytics for Intelligent Healthcare Management

244 CHAPTER 9 INTELLIGENCE-BASED HEALTH RECOMMENDATION SYSTEM

Comparison among MAE and no. of parties
MAE values (patients) MAE values (doctors)

0.935
0.89
0.82 0.79 0.818 0.75 0.69
0.738
MAE value
0.707

1 2 3 4 5

No. of parties
FIG. 9.7
Shows comparison among MAE and number of parties of proposed HRS.

information of their patients and reduces the cost of medical tests by 50%. Cloud-based health infor-
mation searches and intelligent information systems can deliver a highly efficient service model that
covers all the databases of healthcare institutions. It is now easier to improve patient outcomes as well
as illness prevention. Moreover, it helps in the conversion to better and more sophisticated approaches.
This proposed recommendation system in healthcare has also some limitations. As the numbers of users
and items grow, the CF algorithm will suffer serious scalability problems. Another main problem is that
cold-start problem occurs when the HRS does not have enough information about a particular doctor or
patient in order to make relevant predictions. Synonymy is found when a number of the same or very
similar items have different names or entries.

9.6 CONCLUSION AND FUTURE WORK
In this chapter, we exhibit a framework for HRS that works on the basis of big data analysis. It is able to
predict diseases and retrieve data about unknown diseases, which helps in the treatment of patients.
With the input of big data analysis, the whole healthcare system was revived and better solutions to
current healthcare problems can be achieved. With a growing amount of data due to the rising number
of hospital centers and patients all over the world, there is an urgent need to process and analyze un-
structured data. Most public healthcare centers do not follow the procedure in place to avoid unauthor-
ized access to patients’ health records. Centralized patient record systems will become an actuality in
the near future. A centralized server that retrieves and stocks patient health records from different
healthcare centers will lead to easy, hassle-free, and secure data access. Here, different types of big
data analytics techniques are used, which will help in the turnaround and performance of hospital
workers and doctors. Here we focused on three main subareas: clinical data collection, data processing,