Mental Health Assistant using LSTM
0 likes100 views
This document describes a proposed mental health assistant that would use LSTM (long short-term memory) and NLP (natural language processing) to engage in conversation with users, analyze their voice or text input to study their emotion, and provide solutions. It discusses related work on similar mental health chatbots and assistants. The proposed system would take input through either speech or text, use speech recognition and text mining to extract emotion, and provide recommendations to improve the user's mood based on the recognized emotion.
![© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2823
Mental Health Assistant using LSTM
Ms. Aiciri C Hegde1, Ms. Madhiraju Kavyasri2, Ms. Somalaraju Soumya3,Ms. Srinidhi M Sharma4,
Prof. Manohar R5
Students, Department of Information Science and Engineering, Sir M Visvesvaraya Institute of Technology,
Bangalore, Karnataka, India1,2,3,4
Faculty , Department Of Information Science and Engineering, Sir M Visvesvaraya Institute of Technology,
Bangalore, Karnataka, India5
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Mental health is defined as a condition of
well-being in which a person recognises his or her own
skills, can manage with everyday challenges, can work
effectively and can contribute to his or her network. As
the number of people suffering from mental illnesses
rises, it’s difficult to slow or stop the rate of rise in the
number. Our project will engage with the user, gather
data, analyse emotion and provide a solution with the use
of LSTM and NLP. We first begin conversation with the
user then evaluate the voice or text provided by the user
and study the emotion.
Key Words: Mental health, Voice Assistant, Machine
Learning
1.INTRODUCTION
One of the most important and difficult fitness challenges
in the actual world is the automatic identification of
psychological disorders. People’s behaviour, thinking and
emotion are all influenced by their overall mental
wellness,as they engage with the environment. Moreover
mental illnesses are on the rise greatly contributing to the
overall disease load in 2015, it is estimated that 44% of the
people( more than 332 million people) have experienced
symptomsof despair. The condition according to a WHO
research is aprevalent mood disorder that affects a great
amount of people of all ages. There are various barriers to
depressiondiagnosis and therapy along with a paucity of
specialists inthe industry, societal shame and aside-from-
the-point mentality.
1.1 Motivation
Acquiring psychiatric help has been proven to be useful in
the treatment of a variety of health problems particularly
inremote places where mental health facilities are already
scarce. While technological advancements in the field of
mental health received its share of challenges, there were
many who supported online mental health treatments.
Online therapy may not be everyones cup of tea but it has
shown its effectiveness to the people who feel uncomfortable
attending the traditional face-to-face support groups.
Some people show hesitation in sharing their personal
sorrows or problems with otherpeople becauseof the risk of
other people judging them or them getting mocked by other
people. So, in such cases online webpages or dialogue
systems can be used to get the normal informative needs of
the user by acting as a friend or a well-wisher.
1.2 Problem Statement
In the present world of AI and data science, everybody is
searching for the kind of system which is interactive. Often
within the health facility, there may be a loss of human
assets to attend to the patients. So, by means of
considering all these things we have decided our problem
definition as follows: To create a friendly webpage and
integrate it with a voice assistant.
2. RELATED WORK
A. “A Proposal for Virtual Mental Health Assistant”[1]-
Psychological health has been one of the most overlooked
and yet most crucial components of our entire well-being
in recent years. Due to cost, time, and space restrictions, as
well as a scarcity of resources connected to in-person
counselling, this study presents a system for a virtual
mental wellbeing assistant.
Disturbed mental health is frequently the result of a
snowball effect that develops over time and demands
constant attention and purposeful attempts to improve.
With the support of a virtual mental health assistant, this
is achievable. A conversation function, psychological
evaluation, an emotion recognition module, and a
suggestion system for enhancing the user's mood will all
be included in the proposed assistant. For sentiment
analysis, we employed a Naive Bayes classifier and Neural
Networks.
B. “UpBeat – Your Mental Health Assistant using Rasa
NLU”[2]- In
Recent years Textual conversational agent or chatbots
have gained tremendous attention. In today’s world,
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072](https://image.slidesharecdn.com/irjet-v9i6569-221018070757-f340da59/85/Mental-Health-Assistant-using-LSTM-1-320.jpg)
![© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2824
chatbots arewidely used as an agent to communicate with
a human be itas a booking assistant or customer service
support on e- commerce and educational websites. Our aim
of this project is to developed a Cognitive behavioural
therapy system or aTherapy Chatbot which will be able to
assess the mode of the user and also the sentiment of the
interaction that it will have with user. This system or bot
will also devise techniques to provide mindfulness and
provide self-help tothe user so that he/she can take better
care of their life.In this pandemic situation there are
people having emotional breakdowns, some are facing
anxiety, tension,some people even get panic attacks,
depression and face sleep problems. As a result, we are
turning to AI solutions to help these people improve or
maintain their mental health.In a world full of stressful
day-to-day activities no one is completely healthy because
of daily hassles, whether it is employment tension or
feeling of being overwhelmed by other personal factors.
Therefore through this research an effort has beenmade
to facilitate depressive symptoms or aggravation via a
personal virtual psychotherapist.
C. “Mental Health Chatbot System by Using Machine
Learning”[3]-
Mental health has become one of the most overlooked, yet
crucial, aspects of our entire well-being in today's
environment. In this work, we propose a system for a
virtualmental health assistant due to cost, time, and space
constraints, as well as a lack of resources associated with
inperson counselling. Disrupted mental health is typically
the consequence of a snowball effect that necessitates
continual attention and deliberate efforts to remediate.
This is made possible with the help of a virtual mental
health chatbot. The recommended chatbot will have a chat
feature, many language voice input options, and a
recommendation tool to improve the user's mood. Neural
networks were used to train data for this project, and
Natural Language Processing techniques will be used to
improve results.
2. METHODOLOGY
This model can be divided into three steps, the primary
part extracts emotion out of text with the help of text
content mining and LSTMs ,the next part of the application
identifies the key features of speech to recognize the
emotion, in the third part according to the emotion
recognized from the second part solutions are displayed.
Theinputcanbetakenintwo ways- speech or text. We have
set a parameter ,talk and can change the parameter
however we need the input to be.We can set talk as 1 for
speech input and talk as 0 for text input. If text is given as
the input, text mining is used to extract the emotion from
the input. When the input is in speech we use speech
recognition library to convert speech to text, then using
LSTM and NLP the emotion is recognizedand given as the
output.
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072
Fig -1: System Architecture
3. ALGORITHM LSTM Algorithm:
A mapped input example and its output should be present in
the model for the classification procedure. There are a
variety of classification approaches available with Logistic
Regression, Nearest Neighbors, Naïve Bayes, Support
Vector Machines, Decision Trees and Artificial Neural
Networks, being the most commonly employed. The long
short-term memory (LSTM), first established to resolve
the challengesof short-term memory province and permit
the network toconsider longer dependencies. In addition
to the RNN structure, LSTM have a cell state which carries
data through LSTM cells with a simple record exchange as
an outcome. The networks default actions is to take long-
term memory needs into consideration. Furthermore the
LSTM strategy has three gates (forget gate, update gate,
output gate) that indicate the overall set of data that will
be sent to one morecell. The forget gate identifies which
facts ought to be erased. The update gate identifies which
values must reallybe altered and the output gate identifies
the cells output.
Fig -2: Long Short-Term Memory](https://image.slidesharecdn.com/irjet-v9i6569-221018070757-f340da59/85/Mental-Health-Assistant-using-LSTM-2-320.jpg)
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072
Fig -3: LSTM formulas
5. IMPLEMENTATION
The project’s development and deployment is when
aspirations and plans can become actuality. Any software
deployment is always preceded by critical decisions on
theplatform to be utilized, the language to be used and so
on. Several variables have an influence on these choices
including the real-world context in which the system
operates the speed with which it must be implemented,
concerns for safety and other implementation-specific
aspects. It is in reality the programming deployment of a
technical requirements or method as a programme
software development or other computer systems for a
particular standard there may be several
implementations.
4. CONCLUSION
We have created a contextual voice responder which
chatsor responds to the user as per the persons mood. We
havedesigned this voice responder in such manner that it
responds to person if his feeling stressful or lonely and while
he or she is feeling happy. From this we do not want to
give the message that a voice responder will replace an
expert inpsychiatric counselling, this voice responder will
assist person to construct a good mindset toward
lifestyles and breaking down the recursive cycle of
terrible thoughts.
5. FUTURE SCOPE
Our project will chat with a user and tries to understand
hismood and accordingly help him lift up the mood with the
useof LSTM and NLP. The future of this project depends
upon the new developments in the NLP field which will
help in maintaining a proper flow in conversation and all
the dialogues would be automated. We can increase
webpage accuracy by increasing epochs. Epochs are
basically, the number of times your machine learning
model will see eachtraining example during the training.
We can also upgrade the system by recommending the
userconsult the psychiatrists directly through the system.
Also, we can improve the data storage part by recording
the user information and maintaining their records for
future use.
REFERENCES
[1] V. Sreevidya Iyer, Dhanashree Shetty, Purva Badgujar,
Anjali Nambiar, Kajal Jewani, " A Proposal for Virtual
Mental Health Assistant" Published in: 2021 7th
International Conference on Advanced Computing and
Communication Systems
DOI:10.1109/ICACCS51430.2021.9441990.
[2] Aditya Rajesh Nagdekar, Nisha Suresh Poonia, Kaushik
Nayan Dere, Dr.Amol Pande, “UpBeat – Your Mental Health
Assistant using Rasa NLU”, Published in: IRJET Volume: 08
Issue: 05 | May 2021
[3] Prem Patil, Arun Bhau, Shubham Damani, Pritam Pandile,
Prof. Swapnaja Jadhav: “Mental Health Chatbot System by
Using Machine Learning”, Published in: IRJET Volume: 09
Issue: 05 | May 2022
[4] K. Denecke, S. Vaaheesan and A. Arulnathan, "A Mental
Health Chatbot for Regulating Emotions (SERMO) - Concept
and Usability Test," in IEEE Transactions on Emerging
Topics in Computing, doi: 10.1109/TETC.2020.2974478.
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2825](https://image.slidesharecdn.com/irjet-v9i6569-221018070757-f340da59/85/Mental-Health-Assistant-using-LSTM-3-320.jpg)
Mental Health Assistant using LSTM
- 1. © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2823 Mental Health Assistant using LSTM Ms. Aiciri C Hegde1, Ms. Madhiraju Kavyasri2, Ms. Somalaraju Soumya3,Ms. Srinidhi M Sharma4, Prof. Manohar R5 Students, Department of Information Science and Engineering, Sir M Visvesvaraya Institute of Technology, Bangalore, Karnataka, India1,2,3,4 Faculty , Department Of Information Science and Engineering, Sir M Visvesvaraya Institute of Technology, Bangalore, Karnataka, India5 ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Mental health is defined as a condition of well-being in which a person recognises his or her own skills, can manage with everyday challenges, can work effectively and can contribute to his or her network. As the number of people suffering from mental illnesses rises, it’s difficult to slow or stop the rate of rise in the number. Our project will engage with the user, gather data, analyse emotion and provide a solution with the use of LSTM and NLP. We first begin conversation with the user then evaluate the voice or text provided by the user and study the emotion. Key Words: Mental health, Voice Assistant, Machine Learning 1.INTRODUCTION One of the most important and difficult fitness challenges in the actual world is the automatic identification of psychological disorders. People’s behaviour, thinking and emotion are all influenced by their overall mental wellness,as they engage with the environment. Moreover mental illnesses are on the rise greatly contributing to the overall disease load in 2015, it is estimated that 44% of the people( more than 332 million people) have experienced symptomsof despair. The condition according to a WHO research is aprevalent mood disorder that affects a great amount of people of all ages. There are various barriers to depressiondiagnosis and therapy along with a paucity of specialists inthe industry, societal shame and aside-from- the-point mentality. 1.1 Motivation Acquiring psychiatric help has been proven to be useful in the treatment of a variety of health problems particularly inremote places where mental health facilities are already scarce. While technological advancements in the field of mental health received its share of challenges, there were many who supported online mental health treatments. Online therapy may not be everyones cup of tea but it has shown its effectiveness to the people who feel uncomfortable attending the traditional face-to-face support groups. Some people show hesitation in sharing their personal sorrows or problems with otherpeople becauseof the risk of other people judging them or them getting mocked by other people. So, in such cases online webpages or dialogue systems can be used to get the normal informative needs of the user by acting as a friend or a well-wisher. 1.2 Problem Statement In the present world of AI and data science, everybody is searching for the kind of system which is interactive. Often within the health facility, there may be a loss of human assets to attend to the patients. So, by means of considering all these things we have decided our problem definition as follows: To create a friendly webpage and integrate it with a voice assistant. 2. RELATED WORK A. “A Proposal for Virtual Mental Health Assistant”[1]- Psychological health has been one of the most overlooked and yet most crucial components of our entire well-being in recent years. Due to cost, time, and space restrictions, as well as a scarcity of resources connected to in-person counselling, this study presents a system for a virtual mental wellbeing assistant. Disturbed mental health is frequently the result of a snowball effect that develops over time and demands constant attention and purposeful attempts to improve. With the support of a virtual mental health assistant, this is achievable. A conversation function, psychological evaluation, an emotion recognition module, and a suggestion system for enhancing the user's mood will all be included in the proposed assistant. For sentiment analysis, we employed a Naive Bayes classifier and Neural Networks. B. “UpBeat – Your Mental Health Assistant using Rasa NLU”[2]- In Recent years Textual conversational agent or chatbots have gained tremendous attention. In today’s world, International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072
- 2. © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2824 chatbots arewidely used as an agent to communicate with a human be itas a booking assistant or customer service support on e- commerce and educational websites. Our aim of this project is to developed a Cognitive behavioural therapy system or aTherapy Chatbot which will be able to assess the mode of the user and also the sentiment of the interaction that it will have with user. This system or bot will also devise techniques to provide mindfulness and provide self-help tothe user so that he/she can take better care of their life.In this pandemic situation there are people having emotional breakdowns, some are facing anxiety, tension,some people even get panic attacks, depression and face sleep problems. As a result, we are turning to AI solutions to help these people improve or maintain their mental health.In a world full of stressful day-to-day activities no one is completely healthy because of daily hassles, whether it is employment tension or feeling of being overwhelmed by other personal factors. Therefore through this research an effort has beenmade to facilitate depressive symptoms or aggravation via a personal virtual psychotherapist. C. “Mental Health Chatbot System by Using Machine Learning”[3]- Mental health has become one of the most overlooked, yet crucial, aspects of our entire well-being in today's environment. In this work, we propose a system for a virtualmental health assistant due to cost, time, and space constraints, as well as a lack of resources associated with inperson counselling. Disrupted mental health is typically the consequence of a snowball effect that necessitates continual attention and deliberate efforts to remediate. This is made possible with the help of a virtual mental health chatbot. The recommended chatbot will have a chat feature, many language voice input options, and a recommendation tool to improve the user's mood. Neural networks were used to train data for this project, and Natural Language Processing techniques will be used to improve results. 2. METHODOLOGY This model can be divided into three steps, the primary part extracts emotion out of text with the help of text content mining and LSTMs ,the next part of the application identifies the key features of speech to recognize the emotion, in the third part according to the emotion recognized from the second part solutions are displayed. Theinputcanbetakenintwo ways- speech or text. We have set a parameter ,talk and can change the parameter however we need the input to be.We can set talk as 1 for speech input and talk as 0 for text input. If text is given as the input, text mining is used to extract the emotion from the input. When the input is in speech we use speech recognition library to convert speech to text, then using LSTM and NLP the emotion is recognizedand given as the output. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072 Fig -1: System Architecture 3. ALGORITHM LSTM Algorithm: A mapped input example and its output should be present in the model for the classification procedure. There are a variety of classification approaches available with Logistic Regression, Nearest Neighbors, Naïve Bayes, Support Vector Machines, Decision Trees and Artificial Neural Networks, being the most commonly employed. The long short-term memory (LSTM), first established to resolve the challengesof short-term memory province and permit the network toconsider longer dependencies. In addition to the RNN structure, LSTM have a cell state which carries data through LSTM cells with a simple record exchange as an outcome. The networks default actions is to take long- term memory needs into consideration. Furthermore the LSTM strategy has three gates (forget gate, update gate, output gate) that indicate the overall set of data that will be sent to one morecell. The forget gate identifies which facts ought to be erased. The update gate identifies which values must reallybe altered and the output gate identifies the cells output. Fig -2: Long Short-Term Memory
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072 Fig -3: LSTM formulas 5. IMPLEMENTATION The project’s development and deployment is when aspirations and plans can become actuality. Any software deployment is always preceded by critical decisions on theplatform to be utilized, the language to be used and so on. Several variables have an influence on these choices including the real-world context in which the system operates the speed with which it must be implemented, concerns for safety and other implementation-specific aspects. It is in reality the programming deployment of a technical requirements or method as a programme software development or other computer systems for a particular standard there may be several implementations. 4. CONCLUSION We have created a contextual voice responder which chatsor responds to the user as per the persons mood. We havedesigned this voice responder in such manner that it responds to person if his feeling stressful or lonely and while he or she is feeling happy. From this we do not want to give the message that a voice responder will replace an expert inpsychiatric counselling, this voice responder will assist person to construct a good mindset toward lifestyles and breaking down the recursive cycle of terrible thoughts. 5. FUTURE SCOPE Our project will chat with a user and tries to understand hismood and accordingly help him lift up the mood with the useof LSTM and NLP. The future of this project depends upon the new developments in the NLP field which will help in maintaining a proper flow in conversation and all the dialogues would be automated. We can increase webpage accuracy by increasing epochs. Epochs are basically, the number of times your machine learning model will see eachtraining example during the training. We can also upgrade the system by recommending the userconsult the psychiatrists directly through the system. Also, we can improve the data storage part by recording the user information and maintaining their records for future use. REFERENCES [1] V. Sreevidya Iyer, Dhanashree Shetty, Purva Badgujar, Anjali Nambiar, Kajal Jewani, " A Proposal for Virtual Mental Health Assistant" Published in: 2021 7th International Conference on Advanced Computing and Communication Systems DOI:10.1109/ICACCS51430.2021.9441990. [2] Aditya Rajesh Nagdekar, Nisha Suresh Poonia, Kaushik Nayan Dere, Dr.Amol Pande, “UpBeat – Your Mental Health Assistant using Rasa NLU”, Published in: IRJET Volume: 08 Issue: 05 | May 2021 [3] Prem Patil, Arun Bhau, Shubham Damani, Pritam Pandile, Prof. Swapnaja Jadhav: “Mental Health Chatbot System by Using Machine Learning”, Published in: IRJET Volume: 09 Issue: 05 | May 2022 [4] K. Denecke, S. Vaaheesan and A. Arulnathan, "A Mental Health Chatbot for Regulating Emotions (SERMO) - Concept and Usability Test," in IEEE Transactions on Emerging Topics in Computing, doi: 10.1109/TETC.2020.2974478. © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2825