Predicting Mental Health Issues Using Machine Learning Techniques

Kotapati Sai Prasanna, Yagala Suchitra, Talari Umadevi,

Gaddam Drakshayani and Kondakrindi Sucharitha

Department of CSE, Ravindra College of Engineering for Women, Kurnool, Andhra Pradesh, India

Keywords: Mental Health Chatbot, Generative AI, Natural Language Processing, Personality Assessment, Crisis

Detection, Medical Reports, Healthcare Integration, Prompt Engineering.

Abstract: MANAS stands as an advanced mental health chatbot which uses generative AI combined with prompt

engineering together with natural language processing (NLP) to give custom care for people experiencing

mental health problems. MANAS applies a broad personality assessment method to customize dialogues with

users through a combination of their emotional profile and psychological traits made up of openness,

conscientiousness, extraversion, agreeableness, and neuroticism. The system contains crisis detection

mechanisms which can identify suicidal behavior and self-harm so it can activate emergency response

protocols. Through its detailed report generation feature MANAS compiles a summary of patient interactions

as well as emotional states and emerging concerns which healthcare providers can access to support

assessment and treatment. The system resolves current mental health chatbot restrictions through its single

interactive solution that features customized assistance along with distress handling capabilities and

healthcare exchange capabilities. The implementation uses OpenAI API (GPT-3.5) together with NLTK and

spaCy as the underlying library systems.

1 INTRODUCTION

The rising mental health problems which affect

everyone from different backgrounds have become

more widespread throughout the world. Personal

needs of mental health support are higher than

available professionals can manage because of which

patients face extended periods of waiting before

receiving assistance. Traditional therapy and

counseling often require in-person interactions,

which can be inconvenient, costly, and inaccessible for

many. In recent years, digital mental health solutions

have emerged to address these challenges, offering

accessible, cost- effective, and anonymous avenues

for support. However, existing mental health chatbots

often lack the depth of personalization, crisis

detection capabilities, and the ability to integrate

with healthcare systems to provide holistic care. This

gap presents a critical opportunity for innovation in

the realm of mental health technology.

Objectives The MANAS Project (Mental

Assistance Network for Alleviating Suffering) is a

transformative mental health venture based on a

contemporary model that integrates cutting edge

artificial intelligence (AI) with patient-tailored

support methods. The OpenAI GPT-3. 5) MANAS'

simulated human interactions create enriched

servicing for consumers of mental health through the

use of 5 API. The proposed system MANAS has

implemented a sophisticated personality assessment

system, which adapts the response based on

identifying the user's personality, for example, the

behavior of openness and conscientiousness and

introversion and agreeableness and neuroticism. With

this capability the system provides patients with very

specific support that adapts how it is presented to

account for the emotional state of the user.

One aspect of MANAS that sets it apart is its

ability to detect signs of crisis, like suicidal ideation

or self-harm, by analyzing user input. Upon

detecting these vital signals, the behavior-driven

system promptly triggers a pre-designed crisis

protocol to notify appropriate channels, such as

mental health experts or emergency services,

guaranteeing that when users are in critical condition,

they receive prompt assistance. It is an important

move in the direction of an emerging crisis

management strategy which, when properly

equipped-for in advance, stands to save lives through

timely intervention in mental health emergencies.

108

Prasanna, K. S., Suchitra, Y., Umadevi, T., Drakshayani, G. and Suchar itha, K.

Predicting Mental Health Issues Using Machine Learning Techniques.

DOI: 10.5220/0013908900004919

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

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 4, pages

108-114

ISBN: 978-989-758-777-1

With this level of sensitivity, MANAS can facilitate

help not only for users dealing with everyday mental

health and chronic problems but also those seeking

mental health support during acute distress.

Additionally, MANAS extends beyond simple

conversation assistance by creating comprehensive

medical reports summarizing the interaction with the

user, their emotional state, and the possible

problems. These reports are a complete

documentation that can be shared with the health

providers leading to better-advised and quicker

diagnosis. These reports allow users to continue to

receive mental health support, even after they have

engaged in the conversation with the chatbot. The

unique paradigm of easy sharing of medical reports

with professionals makes MANAS an important tool

in a large mental health ecosystem and serves as a

bridge between digital and traditional care.

The new system provides significant updates,

compared to traditional mental health chatbots,

because it can maintain full medical records, offer

customized conversations and assist in crisis

situations.

The MANAS system merges multiple services

using composite personality assessment and crisis

detection capabilities along with report generation

functions to assist the individuals dealing with

mental health challenges. This great system has also

got some AI-powered dialog and gives you a

personalized interaction at your fingertips that

supports you throughout every interaction.

What makes the MANAS Project innovative is its

integrative approach to mental health care, bringing

together state-of-the-art AI technology with a rich

understanding of human psychology. MANAS

aspires to provide a safe and healthy space for users

to seek help around mental health challenges by

processing emotional cues and tracking responses

accordingly in a direction that is gentle, caring and

empathic. The primary goal of MANAS is not only to

provide immediate support services but also to

connect them to longer-term mental health care for

sustained health and recovery.

2 LITERATURE SURVEY

The advent of artificial intelligence (AI) has played a

part in revolutionising mental health care worldwide.

Olawade et al. (2024) examines the current trends and

the future prospects of optimizing mental health with

artificial intelligence, detailing its transformative

potential in the diagnosis and treatment of mental

health issues. AI-Powered Solutions for Mental

Health CareAI-enabled solutions address the mental

health gap with immediate, customized solutions. The

Ajayi (Singh, L. (2024) study proposes the use of AI

driven technology as an asset to manage complex

mental health crisis and also towards addiction

treatment programs and is a sign of AI's ability to

provide targeted treatment strategies at scale.

AI applications for specific uses in mental health

have received attention in several investigations in

the last few years. Lee et al. analyze the clinical

applications of AI in mental health care.

(Koutsouleris et., al. 2022) that address both

diagnostic aid and treatment capabilities, as well as

challenges that concern ethics and clinical uptake.

Koutsouleris et al. We discuss moving the promises

of AI into a functional mental healthcare

infrastructure (Thirupathi et., al. 2025) for real-world

use within clinical practice. This research illustrates

the potential opportunities and challenges that the

introduction of AI systems poses for the delivery of

mental health services.

Singh introduces the concept of AI-human

synergy in mental health, emphasizing the potential

for creating mental health applications that support

mental health workers' well-being. This approach

explores how AI can serve as a supportive tool for

professionals, enhancing their ability to manage

complex cases. On the other hand, Thirupathi et al.

2025 investigate the role of AI and the Internet of

Things (IoT) in providing continuous, personalized

mental health care, offering solutions for digital

diagnostics and long-term support. Their research

points to the evolution of mental health tools from

crisis management to ongoing, individualized care.

AI through chatbots presents itself as a valuable

solution for mental health support when built

specifically for this purpose. The study by Yoon

examines AI-based digital therapeutics that handle

adolescent mental health care programs focusing on

disaster response applications. The research by

Manole et al. 2024 features chatbots which use AI for

anxiety intervention through individualized mental

health assistance Dutta, D., & Muni, A. D. (2024).

The applications serve both urgent emotional

requirements as well as establishing continuous

programs for mental health maintenance.

Scientific research includes AI usage in mental

health fields through its application to psychiatric and

nursing care systems. Nashwan et al. examine how AI

strategies enhance psychiatric patient care by

explaining its value for mental health nursing practice

to reach better outcomes. According to Dutta and

Muni (2024), AI demonstrates its ability to detect

depression early alongside managing this condition

Predicting Mental Health Issues Using Machine Learning Techniques

109

within elderly populations.

The healthcare field is now tackling issues of

security and morality that arise from AI

implementation in mental health practices. De Freitas

et al. (2024) present essential information about the

security aspects of generative AI chatbots for mental

health alongside necessary precautions for their

appropriate utilization. Singh (2023) explains that AI

tools such as ChatGPT will reshape mental health care

but healthcare providers need to supervise their

implementation to maintain ethical compliance. The

evaluation by Denecke et al. (2021) examines the

implementation risks and opportunities of AI-

powered chatbots in mental health care while

identifying the requirement for safe and effective

methods of deployment.

Two research studies by Ahmad et al. (2022) and

Sweeney et al. (2021) focus on personality-adaptive

conversational agents designed for mental health care

use. The research by Ahmad et al. discusses how

agents responding through conversation should adjust

to precisely match users' personalities to maximize

their impact. This study by Sweeney et al. investigates

how healthcare professionals perceive chatbots

through their evaluation of potential advantages along

with integration barriers for AI in mental health

service delivery.

3 PROPOSED METHODOLOGY

The MANAS Project provides services to mental

health patients by integrating artificial intelligence

generative AI solutions with natural language

processing and personality evaluations. By

identifying crisis moments as well as delivering

tailored ongoing patient care, the system delivers a

full healthcare services package. The system provides

empathetic and accurate answers because it recognize

user emotional states along with personality features

and environmental elements. The structure presented

consists of essential portions that initially capture

information and employ man made intelligence

calculations. Figure 1 shows the system architecture.

3.1 Personality Assessment and

Emotional Context Detection

MANAS begins the process with personality

evaluation based on the Five-Factor Model (FFM) to

analyze users traits of openness, conscientiousness

extraversion, friendliness and neuroticism. Detecting

emotional cues in user inputs, analyzing the

emotional states, and creating a database of emotional

cues leads to an assessment of user emotions. The

necessary knowledge to adjust conversation style is

provided to the system by user feedback, as it allows

the chatbot to offer suggestions that align with a

person's traits. The personalized response system

builds emotional resonance with users who face

mental health challenges creating an emotionally

nuanced tone for mental health care support. At the

same time, MANAS also uses state-of-the-art NLP

techniques to recognize the emotional context of the

user's input and detect the signs of distress, sadness,

anxiety or other emotional state. This information is

essential to modify the chatbot response and tone

depending on real-time user interaction. For instance,

if the system senses that a user is feeling low or

stressed, the chatbot might take a gentler, more

encouraging approach. Detecting the emotional

context is essential for improving user interaction,

establishing trust and ensuring the chatbot responds

ideal to different emotions.

3.2 Crisis Detection and Emergency

Protocol Activation

The second segment of the methodology uses

intelligent algorithms to identify signs of a crisis

through analytical of user- generated text contents for

recognized warning signs. The model is extensively

trained on diverse mental health datasets to learn the

patterns of identification for speech content related to

crisis and mental health keywords. Once a mental

health crisis is detected by the system it employs a

predetermined alert protocol to reach out to relevant

stakeholders such as crisis hotlines along with mental

health professionals and other emergency contacts.

MANAS has a distinctive approach where a ready-

made model of a conversation integrates users with

instant help which is its competitive advantage

against other mental health chatbots.

We also propose an escalation mechanism that

adjusts intervention severity according to the type of

crisis detected, thus further refining the crisis

detection process. For instance, if a user is reported

to be suicidal, the MANAS can prioritize responding

quicker by calling emergency services or send a

message to the user's primary emergency contact. The

goal of the system is to respond based on urgency

and not leave the user in a situation where they are

dealing with a life-threatening situation alone.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

110

3.3 Data Collection and Continuous

Care through Medical Reporting

The third component of the methodology is the

ongoing collection and analysis of user data for

mental health monitoring. As users engage with the

chatbot, their emotional state, responses, and

behavioral patterns are tracked for analysis. These

logs are not personally identifiable, but they give us

valuable information about how the user interacts

with us. The system analyzes the collected data and

automatically generates detailed medical reports that

summarize the user's emotional state, behavioral

trends, and interactions. These reports provide

medical professionals with context to evaluate the

mental state of the user over time.

MANAS delivers ongoing treatment by

providing users with a platform to monitor their

mental health journey. A user can show these reports

to healthcare providers, enabling continuous mental

health support and better clinical assessments. It

provides an objective perspective of the user’s mental

health trajectory, and supplements physical therapy

approaches by documenting real-time indicators of a

user’s mental state and emotional excesses.

3.4 Integration and User Experience

Lastly, the methodology emphasizes on the inclusion

of interface that allows the user to interact with

system and access all other components in a unified

way. Our chatbot is user-friendly with which to

communicate so it offers intuitive options for to start

conversations, track their feelings and ask for

assistance. You are built on a modern tech stack using

ReactJS for the front-end, providing an excellent

user experience on all devices. It is pivotal that user

interface shall go hand in hand with AI-driven

algorithms for real-time personalized support, which

makes the experience better for the user and

motivates them for continuous interaction with the

system.

In addition, MANAS acts to bridge a gap in the

human web of healthcare professionals, tapping them

into existing healthcare systems for easy transfers

when users need it. This integration uses secure

sharing features to let healthcare providers view user

reports and evaluate their mental health alongside

other clinical data. MANAS bridges the gap between

digital health interventions and traditional healthcare

systems by facilitating a holistic, continuous care

process, where users can access professional help on

demand when they need it most.

Figure 1: System architecture.

4 RESULTS AND DISCUSSION

The MANAS Project uses machine learning, natural

language processing and computer graphics to create

a highly personalized and empathetic mental health

support system. These methods will yield anticipated

results for both the user experience and mental health

care: The system harnesses complex data pattern

extraction to create profiles of users' emotional

states, behavioral patterns, and contextual

information, so as to effectively present tailored

responses. One of the most important functions of the

chatbot is its ability to adapt to different

personalities, emotional signals, and changes in

external circumstances, such as crisis detection, to

allow the system to respond in dynamic and

personalized ways to users.

We expect the model to adapt to a variety of user

inputs, making it more intuitive and empathetic. For

example, during periods of heightened emotional

distress, the system is expected to change the tone and

topic of its content to deliver more supportive,

comforting and urgent responses. Moreover, if users

display signs of a mental health crisis, such as suicidal

thoughts, the system will trigger appropriate

emergency responses, notifying healthcare

professionals or crisis hotlines instantaneously - all

throughout the process, personal information and

concerns will never be stored., this adaptable model

will learn to recognize and respond to different

emotional states and provide a more effective,

targeted solution than has historically been possible.

4.1 Expected Trends in User

Interactions

The user will most likely have different behavior and

Predicting Mental Health Issues Using Machine Learning Techniques

111

data than other users, which the model should pick

up on. For instance, users with high rates of

neuroticism might show more frequent signs of

emotional distress and may warrant more immediate

crisis detection and intervention. On the other hand,

users who tend to show more agreeableness or

extraversion are likely to have more casual and

positive conversations, making space for long-term

mental health interventions for the chatbot rather

than immediate intervention. Therefore, the system's

flexibility is needed to cope with these dynamics in

real-time, providing personalized and relevant

responses grounded on the user's individual

emotional and psychological features.

Figure 2: Response time and engagement level based on

personality traits.

Figure 3: - This Graph Shows What Predicts Different

Personality Traits (Neuroticism, Extraversion, Etc) As Well

As the Time It Takes The Chatbot to Respond (In Minutes)

and How Engaged the User Was (Score of 1-10). Figure 2

Shows the Response Time Is Depicted Through the Red

Bar, and Engagement Levels Through The Blue.

Other anticipated developments are the system’s

capacity to observe trends in people’s responses

throughout time, allowing it to monitor mental health

progress and adjust support accordingly. For

example, if a user shows improved mood and

emotional well-being, the system may provide

suggestions for their continued mental health journey,

whereas a user who returns with a decline in their

emotional well-being may be offered an immediate

intervention or referred to professional care. Table 1

shows impacts on user behaviour.

4.2 Comparative Analysis: MANAS vs.

Traditional Methods

Comparative to traditional therapy which is time

consuming, expensive and requires physical

attendance, MANAS gearing up better advantages by

providing standard mental health service provision.

MANAS provides users constant availability as well

as personalized representation support which also

captures and stores user progress independently over

time. Instant therapeutic solutions for users via AI-

based operational models housing these that extract

mental health support through emotional input and

identify crises at this very moment.

Traditional approaches may often return profuse

examples of querying and information which do not

offer the same level of relevance depending on the

information received from the user; the MANAS

system works to differentise itself in this aspect. High

user engagement and user satisfaction compared

with standard mental health chatbots without

personalization is likely to be an outcome of this.

Additionally, the incorporation of crisis detection and

medical report generation offers a holistic, end-to-end

solution that is unparalleled by traditional systems.

Figure 3: Expected behaviour patterns based on personality

types.

Table 1: Expected impact of external data (weather,

economic variables) on user behavior.

External

Factor

Expected

Impact on

User Behavio

Notes

Weather

(Seasonal)

Increased

depressive

symptoms in

colder

months.

Emotional cues

might indicate a

need for more

support during

winter.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

112

Economic

Conditions

Higher stress

levels due to

economic

downturns.

System may detect

increased anxiety

and offer targeted

interventions.

Social Events Positive mood

during

holidays,

higher

ement.

The system may

shift its tone to

reflect a more

positive

environment.

Media

Exposure

Increased

exposure to

negative news

could cause

distress.

Adjustments in

response tone, with

crisis detection if

needed.

5 CONCLUSIONS

MANAS Project is an important developmental step

in digitalmental health support due to its integration

of generative AI technologywith a personality

assessment and crisis detection in a single interface.

Through analysis of thought and act in emotional

state and personality traits, MANAS can generate

real-time responses allowing MANAS to provide

mental healthcare with a higher degree of empathy

compared to the traditional approach. The system is

capable of detecting crises and issuing complete

medical records for continuous health care needs,

thus enabling patients to achieve complete mental

wellness from the entire system. Mental health

support systems face a paradigm shift for the future

as it will provide better support methods that suit the

individual needs as well as support operational

outputs to improve therapeutic outcomes on the

mental health of a patient.

6 FUTURE SCOPE

Capabilities that support the mental health

component of the MANAS Project will have the

highest leverage impact. Recommended Future Work

Further improvement of MANAS Project could user

multi patient data as Vocal tone, biological sensor

data for complete emotional state analysis. In

addition, the model can be expanded in its

applications by integrating multiple health conditions

along with assessing them continuously through

wearable sensors so as to enhance its usability in pre-

crisis and crisis realms of treatment. Improving

Multilingual Support Features The system should be

able to address a broader range of language

populations that it can serve. Incorporation with

Electronic Health Records (EHRs) and telemedicine

systems enabling seamless communication between

traditional and digital mental health services will

make MANAS further integrated with healthcare.

REFERENCES

Ahmad, R., Siemon, D., Gnewuch, U., & Robra-Bissantz,

S. (2022). Designing personality-adaptive

conversational agents for mental health care.

Information Systems Frontiers, 24(3), 923-943

De Freitas, J., Uğuralp, A. K., Oğuz‐ Uğuralp, Z., &

Puntoni, S. (2024). Chatbots and mental health: Insights

into the safety of generative AI. Journal of Consumer

Psychology, 34(3), 481-491.

Denecke, K., Abd-Alrazaq, A., & Househ, M. (2021).

Artificial intelligence for chatbots in mental health:

Opportunities and challenges. In Multiple Perspectives

on Artificial Intelligence in Healthcare: Opportunities

and Challenges (pp. 115-128).

Depression in Elderly Citizens. Library of Progress-Library

Science, Information Technology & Computer, 44(3).

Dutta, D., & Muni, A. D. (2024). Artificial Intelligence for

Mental Health: A Review Based Analysis of Early

Detection and Management of

Koutsouleris, N., Hauser, T. U., Skvortsova, V., & De

Choudhury, M. (2022). From promise to practice:

towards the realisation of AI-informed mental health

care. The Lancet Digital Health, 4(11), e829-e840.

Manole, A., Cârciumaru, R., Brînzaș, R., & Manole, F.

(2024). Harnessing AI in Anxiety Management: A

Chatbot-Based Intervention for Personalized Mental

Health Support. Information, 15(12), 768.

Nashwan, A. J., Gharib, S., Alhadidi, M., El-Ashry, A. M.,

Alamgir, A., Al-Hassan, M., ... & Abufarsakh, B.

(2023). Harnessing artificial intelligence: strategies for

mental health nurses in optimizing psychiatric patient

care. Issues in Mental Health Nursing, 44(10), 1020-

1034.

Olawade, D. B., Wada, O. Z., Odetayo, A., David-

Olawade, A. C., Asaolu, F., & Eberhardt, J. (2024).

Enhancing mental health with Artificial (AI) Graham,

S. A., Kim, H. C., ... & Jeste, D. V. (2021). Artificial

intelligence for mental health care: clinical

applications, barriers, facilitators, and artificial

wisdom. Biological Psychiatry: Cognitive

Neuroscience and Neuroimaging, 6(9), 856-864..

Singh, L. (2024). Revolutionizing Care: Developing an AI-

Human Synergy Mental Health Application for Mental

Health Worker’s Well-Being (Doctoral dissertation,

Capella University).

Singh, O. P. (2023). Artificial intelligence in the era of

ChatGPT: Opportunities and challenges in mental

health care. Indian Journal of Psychiatry, 65(3), 297-

298.

Sweeney, C., Potts, C., Ennis, E., Bond, R., Mulvenna, M.

D., O’Neill, S., ... & McTear, M. F. (2021). Can

chatbots help support a person’s mental health?

Perceptions and views from mental healthcare

Predicting Mental Health Issues Using Machine Learning Techniques

113

professionals and experts. ACM Transactions on

Computing for Healthcare, 2(3), 1-15.

Thirupathi, L., Kaashipaka, V., Dhanaraju, M., &

Katakam,V. (2025). AI and IoT in Mental Health Care:

From Digital Diagnostics to Personalized, Continuous

Support. In Intelligent Systems and IoT Applications in

Clinical Health (pp. 271-294). IGI Global.

Yoon, S. (2024). AI-Based Digital Therapeutics for

Adolescent Mental Health Management and Disaster

Response. Information, 15(10), 620.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

114