Analysis of the Functions of Database Information Management

System: Taking the User Health Record Platform as an Example

Ying Xiang

Aberdeen Institute of Data Science and Artificial Intelligence, South China Normal University,

Taoyuan East Road, Foshan City, China

Keywords: Database, Data Security, Data Prediction.

Abstract: Mental illness is recognized as an important public health problem globally. The disease is accompanied by

a huge number of complications and often masks the real psychological problems, causing a lot of losses in

manpower, material and financial resources. And currently, medical data often has biases and imbalances.

Certain types of case data are sparse, which can easily lead to errors in prediction and diagnosis by algorithms.

At the same time, it is important to ensure that patients provide personal data and protect privacy rights. This

article concludes that the use of databases can assist doctors to accurately diagnose the corresponding mental

illness from a large number of complications faster and provide timely treatment. This article mainly uses a

relational database based on the basic functions of the database to conduct research on security and prediction

accuracy. Through discussion, this article combines access control and Bayesian networks with databases to

make the prediction results in practical applications safer and more accurate and provides relevant ideas for

combining databases with disease research.

1 INTRODUCTION

Mental illness is considered an important public

health problem globally, with a huge number of

accompanying complications that often mask the real

psychological problems, causing a lot of losses in

human, material and financial resources. As an

efficient data management system, the database can

store a large amount of relevant data and predict new

situations based on past data. Therefore, database

management can provide researchers with

comprehensive and accurate case records and

treatment data, helping to reveal the characteristics

and evolution of the disease.

The database not only provides efficient data

management but can also be used to store patients'

personal information, past medical records, treatment

plans, treatment results and other related information.

Through the database, medical staff can further

analyze the patient by sorting the data to ensure

accurate treatment and predict the possibility of

recurrence of the disease, thereby making targeted

recommendations. At the same time, medical staff

can also search for similar cases based on keywords

https://orcid.org/0009-0000-6036-7064

in order to use past successful cases and experience

to treat new patients accordingly, reducing the

possibility of incorrect diagnosis and improving

diagnosis and treatment efficiency. Healthcare

professionals also need to take data security seriously,

especially when handle personal patient information,

to ensure the accuracy and trustworthiness of research

on mental illness. However, in the era of big data,

ensuring the security of patients' personal data and

protecting the privacy of users is particularly

important. Moreover, current medical data often have

biases and imbalances, and there are less data on

certain types of cases, which can easily lead to errors

in prediction and diagnosis by algorithms, resulting in

lower credibility of prediction results. Therefore, it is

particularly important to build a database with high

security performance and accurate prediction

capabilities.

In order to predict mental illness more accurately,

the international cooperation organization Cochrane

Collaboration published a medical literature database

called Cochrane Library, which contains a large

number of systematic reviews of clinical trials of

mental illness; the American Psychological

Xiang, Y.

Analysis of the Functions of Database Information Management System: Taking the User Health Record Platform as an Example.

DOI: 10.5220/0012916100004508

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence (EMITI 2024), pages 159-163

ISBN: 978-989-758-713-9

159

Association (APA) operates a database called

PsycINFO The psychology literature search database

contains a large number of studies on mental illness.

These databases provide valuable support and

reference for mental illness research. At the same

time, some hospitals have also adopted related

decision-making systems, such as decision trees,

logistic regression support, etc., to establish platforms

to screen conditions. However, due to the

characteristics of mental illness and the current low

popularity of mental illness-related knowledge, there

is less reliable data available. Therefore, the research

on the combination of this field with databases is

relatively blank and has high research value.

This article explores how databases can be

designed and used to address this issue and ensure the

protection of patient privacy. This article focuses on

the functions and functions of the database. Then, this

article mainly analyzes how to improve the security

of the database and the accuracy of diagnosis and

treatment of mental illness through reasonable

algorithms and methods.

2 BASIC FUNCTION

INTRODUCTION

A database is an important tool for storing data. It has

a variety of functions that can adapt to the needs of

different application scenarios. The main functions of

the database include data storage, data retrieval, data

security, data analysis and reporting.

Data storage is the basic function of the database.

Organize corresponding information by establishing

tables and inserting data to store structured and

unstructured data orderly. Databases can also store a

variety of data, such as numbers, text, dates, images,

audio and video. At the same time, database

administrators can complete data insertion, update,

and deletion operations, either individually or in

batches, to meet different business needs. In addition,

the database can also ensure data consistency through

transaction management and ensure data integrity in

multiple operations.

Secondly, the database has powerful data retrieval

capabilities. Users can use query languages such as

SQL to filter, sort, group, and other data based on

specific conditions to quickly obtain the required

information and facilitate data use and extraction.

The database also provides a variety of

mechanisms to ensure data security, such as access

control, user authentication, and data encryption. By

setting permissions, users can restrict their access to

and operations on data, prevent unauthorized access

and data leakage, and facilitate data management. At

the same time, databases can use advanced encryption

techniques such as secure hashing algorithms (SHA)

to generate unique identifiers in a set of variables, and

by including a reasonable number of variables to

identify unique patterns, it is possible to compare

different data sets. Generate an anonymous unique

identifier and merge them (Marco et al., 2023). In

addition, the database can back up data regularly to

prevent data loss, and in the event of data damage or

loss, the data can be restored to its most recent state

through the recovery function to prevent data loss to

the greatest extent.

Moreover, using the easy-to-understand OLAP

paradigm for multi-dimensional big data analysis, and

through its flexibility and expressive capabilities,

helps people discover the relationship of data from

different angles and analyze the data to help users

better understand the data and make better decisions.

decision-making (Alfredo, 2023). At the same time,

the database can also predict future data based on

other algorithms, allowing people to take better

measures for the future or improve current functions.

To sum up, the database, as a powerful data

management tool, has a variety of functions to meet

the needs of different application scenarios. With the

development of technology, the functions of the

database are constantly enhanced and improved,

providing better support for various application

scenarios.

3 SPECIFIC FUNCTION

INTRODUCTION

3.1 Data Security

With the rapid development of Internet technology

and the popularity of network applications, the

number of websites has shown explosive growth.

These websites cover various fields, including but not

limited to government departments, commercial

enterprises, educational institutions, medical

institutions and various social groups. They have built

information exchange platforms by establishing

websites to provide users with various services and

information. Database systems play a vital role

behind the scenes of these websites. They are

responsible for storing and managing large amounts

of user data, including personal information,

transaction records, social interactions, etc. However,

with the surge in the number of websites and the

EMITI 2024 - International Conference on Engineering Management, Information Technology and Intelligence

160

expansion of data volume, ensuring that these

websites can run safely and stably has become a

major challenge for website administrators and

database administrators.

The complexity and changeability of the network

environment, the numerous loopholes in information

systems, and the increasingly sophisticated hacker

attack methods have put database security protection

under tremendous pressure. In the design of

databases, most databases use commercial database

systems such as DB2 and SQL Server. Although these

database systems have mature technical conditions,

problems such as SQL injection may still occur.

However, considering the stability of the database,

the update speed of more complete patches is usually

delayed, which ultimately makes it difficult to deal

with database vulnerabilities the first time, and

eventually becomes a security risk (Peng et al., 2022).

Hackers often use system vulnerabilities to steal

confidential data or destroy important data by

invading databases to achieve their illegal purposes.

This not only violates users' privacy rights, but also

brings huge economic losses and reputational damage

to website operators.

In order to deal with these security threats,

database administrators have taken a variety of

measures to protect data security, such as setting

access control, data encryption, auditing and

monitoring functions, SQL injection protection, etc.,

so that even if the data is illegally obtained, It also

cannot be easily interpreted, preventing hackers from

exploiting these vulnerabilities. At the same time,

detection technology based on machine learning can

also be used to train a machine learning model, so that

the machine can distinguish between normal SQL

statements and malicious SQL statements by learning

algorithms such as decision trees, support vector

machines, and neural networks, so that it can

automatically Identify patterns of malicious behavior.

Although this detection method requires a large

amount of training data and computing resources, it

can detect unknown attack methods and has high

detection accuracy (Zhu et al., 2024).

In addition to the above measures, the database

server is also protected by physical security. Such as

installing surveillance cameras, using access control

systems, establishing security barriers, etc., to protect

the physical location of the database server and

prevent unauthorized physical access.

In short, with the continuous development of

network applications, database security has become

an important issue that cannot be ignored. Through

the comprehensive use of technology and

management methods, the security of the database

can be greatly improved, user data is protected from

infringement, and the normal operation of the

database is maintained.

3.2 Data Prediction

The prediction function of the database is based on

existing data and uses data analysis and machine

learning algorithms to predict future trends. Its steps

mainly include data collection and cleaning, data

analysis and modeling, model evaluation and

optimization, and prediction and application. Human

society generates a large amount of different types of

data every day with people's social activities. If these

large data sets are to be useful, new tools are needed

to analyze them, which has led to the application of

machine learning (ML). Liu et al. (2020) define ML

methods as the practice of generating predictive

models based on learning feature patterns from data,

and predicting new data or results through the

constructed models (Flores et al., 2023).

In addition, transfer learning and adding

explainable artificial intelligence can also help

database predictions become more accurate. Transfer

learning can fully use the feature representation

capabilities of existing models in new disease

prediction by utilizing existing data and models, such

as artificial learning network models, thereby

improving the accuracy and generalization

capabilities of the new model. Migration learning has

a high data utilization rate. It can extract and use

existing models' feature extraction capabilities to

greatly reduce data analysis and annotation and

improve data utilization. At the same time, it can

avoid building a model from scratch. However,

suppose there is a large difference between the

original rules and the new rules. In that case, the effect

of transfer learning may decrease, and some

irrelevant features may be introduced, which may

reduce the model's performance.

In actual use, while ensuring the reliability of

predicted data, it is also necessary to provide users

with a clear decision-making process, so explainable

artificial intelligence needs to be added. Designers

can use the SHAP model in game theory: Shapley

value. In a game, participants cooperate to achieve

certain benefits. The purpose of the Shapley value is

to allocate each player's contribution to different

ways of cooperation. For example, the Shapley value

defines a reasonable distribution mechanism so that

the benefits obtained by each participant are

proportional to their contribution to the game

(Lundberg et al., 2017). Database administrators can

use this method to determine the data type that has a

Analysis of the Functions of Database Information Management System: Taking the User Health Record Platform as an Example

161

greater impact on the final result and quickly find the

corresponding relationship.

4 DATABASE DESIGN: TAKING

THE MENTAL ILLNESS

RECORDING PLATFORM AS

AN EXAMPLE

4.1 Method of Data Storage

Usually, data storage methods are divided into two

types of databases: relational and non-relational.

Since different complications of patients have

different characteristics, this article chose a relational

database as the data storage method for this database.

Relational databases can use relational algebra and

powerful Structured Query Language (SQL)

capabilities to filter information within the database,

allowing doctors to solve complex queries through

combined operations (Córdoba-Hidalgo et al., 2022).

Therefore, doctors can use a patient's symptoms to

search the database to see if other patients also have

the same symptoms and find the best medical advice

and treatment while identifying possible

complications. Through this method, doctors can

quickly locate the patient's symptoms, screen out

other hidden diseases, strive for the best treatment

time, and help the patient recover quickly.

4.2 How to Ensure Data Security

SQL injection attack is a relatively common database

attack method. It is mainly based on the SQL

language database system and uses vulnerabilities in

web applications to inject attacks on the database to

create, read, modify or delete sensitive data (Kasim,

2021). Since vulnerabilities in Web systems are

unavoidable, Text Plan adopts access control to

formulate policies based on the principle of least

privilege to prevent relevant information from being

directly exposed to users in order to deal with security

threats such as SQL injection attacks and ensure the

security of the database. This article uses role-based

access control to classify different users, grant

different permissions, and manage permissions. For

example, doctors can query patient information and

apply it to submit new case information, and only

database administrators can add, modify, or delete

cases. At the same time, when users log in, they also

need to pass user authentication, such as entering user

name, password, using multi-factor authentication,

etc. to ensure that the identity of the login is authentic

and reliable, so as to ensure the security of the

database.

4.3 Data Prediction

By using Bayesian models for probabilistic inference,

the impact of other diseases can be reduced and the

condition can be determined more accurately.

Bayesian network is a graphical model used to

represent dependencies between variables and

provide explainable results through probabilistic

inference (Koller et al., 2009). Bayesian networks can

use the patient's symptoms, signs and examination

results as nodes, various possible diseases as output

nodes, and the dependencies between nodes are

expressed as directed edges. The probability of each

disease is provided by calculating the posterior

probability, thereby providing prediction results for

the patient's disease.

Patients often have more than one symptom, so a

multi-feature fusion method based on Bayesian

theory is applied. The performance and generalization

ability of the model can be improved through feature

combination, feature selection, etc., and the results

can be further accurately predicted (Li, 2019).

Feature combination is to combine different

features to generate new features. For example, a

patient's two physical conditions, A and B, can be

multiplied to obtain a new feature, AB. This captures

the interaction between features and provides more

information for analysis on the Bayesian network.

Feature selection selects the most representative

and effective features among multiple features and

assigns an important value or weight to the features

for feature ranking (Kanti et al., 2021). People can use

methods such as correlation, information gain, and

chi-square tests to evaluate the degree of association

between features and target variables, and select

features with higher correlations for analysis to

reduce negative impacts. For example, by assessing

the correlation between the patient's various physical

indicators and the current discomfort and by selecting

highly relevant indicators, people can more quickly

find the potential cause and provide rapid diagnosis

and treatment.

5 CONCLUSIONS

This article conducts research and analysis on the

functions of the database, protecting its security and

increasing the accuracy of data prediction. Taking the

database design of the mental illness recording

platform as an example, a relational database is used

EMITI 2024 - International Conference on Engineering Management, Information Technology and Intelligence

162

to facilitate users to quickly search for relevant

information. At the same time, access control is used

to prevent SQL injection attacks and reduce the

exposure of key information to users to ensure

database security. In terms of prediction, Bayesian

models are used to conduct correlation analysis on

various complications, and explainable artificial

intelligence is added to facilitate doctors to confirm

patients' symptoms faster. By calculating the

probability to find the main cause of the disease, the

accuracy and reliability of the prediction results are

ensured while excluding complications and other

erroneous data.

Although the Bayesian network assumes that each

condition is independent, the model is relatively

simple and computationally efficient and can avoid

erroneous data through probabilistic analysis.

However because of this assumption, the model

struggles to make complete predictions. Moreover,

this model requires a large number of training

samples for data training, and there is still a lack of

samples in this field, so it is still difficult for this

model to be accurately applied in disease diagnosis.

Due to the small number of relevant references, it

is difficult for this article to form an accurate

feasibility analysis conclusion. At the same time, due

to the lack of detailed understanding of the human

nervous system and convolutional neural networks, it

is still at a highly theoretical stage and is difficult to

apply in specific practice. However, the Bayesian

network proposed in this article can be used to

analyze new cases in the database, provide ideas for

subsequent research, and reduce the severe

psychological burden on patients due to the impact of

long-term complications and further aggravate

mental illness.

REFERENCES

Alfredo, C,2023, Compressing Big OLAP Data Cubes in

Big Data Analytics Systems: New Paradigms, a

Reference Architecture, and Future Research

Perspectives. Communications in Computer and

Information Science, 1849, CClS.

Córdoba-Hidalgo, P., Marín, N., & Sánchez, D,2022, RL-

instances: An alternative to conjunctive fuzzy sets of

tuples for flexible querying in relational databases.

Fuzzy Sets and Systems, 445, 184–206.

Flores, K. R., de Carvalho, L. V. F. M., Reading, B. J.,

Fahrenholz, A., Ferket, P. R., & Grimes, J. L,2023,

Machine learning and data mining methodology to

predict nominal and numeric performance body weight

values using Large White male turkey data sets. Journal

of Applied Poultry Research, 32(4), 100366.

Kanti Ghosh, K., Begum, S., Sardar, A., Adhikary, S.,

Ghosh, M., Kumar, M., & Sarkar, R,2021, Theoretical

and empirical analysis of filter ranking methods:

Experimental study on benchmark DNA microarray

data. Expert Systems with Applications, 169, 114485.

Kasim, Ö,2021, An ensemble classification-based approach

to detect attack level of SQL injections. Journal of

Information Security and Applications, 59, 102852.

Koller, D., & Friedman, N,2009, Probabilistic graphical

models: Principles and techniques. The MIT Press

eBooks.

Li, Z. H,2019, To explore the underlying molecular

mechanism of mental illness based on the comorbidity

and multi-feature fusion Bayesian model [Doctoral

dissertation, East China Normal University]. Retrieved

from

https://kns.cnki.net/kcms2/article/abstract?v=0Vs2Vpq

j5wcperIKNbCrv7xQoET3WbTi_jdtg7TSc0X5bbDgu

OSggsDT8Mw9AJR_LOcfcdmEkWc2io7BTO4vA_d

SqxFIsWj05h5KGpIAb2f9JX1cGunKAR85nPorN_O

wG_1uX5Apb5U=&uniplatform=NZKPT&language=

CHS

Lundberg, S., & Lee, S.-I,2017, A Unified Approach to

Interpreting Model Predictions. arXiv: Artificial

Intelligence.

Peng, X. M., & Hu, J,2022, Database Security System

under the background of big data. Computer knowledge

and technology, 09, 11-12.

Tomietto, M., McGill, A., & Kiernan, M. D,2023,

Implementing an electronic public health record for

policy planning in the UK military sector: Validation of

a secure hashing algorithm. Heliyon, e16116.

Wang, F,2019, Digital Communications World. Journal of

Digital Communication, 25(4), 256-268.

Wang, Z,2023, Modern industrial economy and

informatization. Journal of Modern Industrial Economy,

35(1), 32-45.

Zhu, Z. N., & Jin, J. Q,2024, Research on attack technology

based on SQL Injection vulnerability. Computer

knowledge and technology, 01, 98-100 + 103.

Analysis of the Functions of Database Information Management System: Taking the User Health Record Platform as an Example

163