Real Time Detection of Human Stress using Sensors and Machine

Learning Techniques

Abstract
Stress may be defined as the reaction of the body to regulate itself to
changes within the environment through mental, physical, or emotional
responses. Recurrent episodes of acute stress can disturb the physical and mental
stability of a person. This subsequently can have a negative effect on work
performance and in the long term can increase the risk of physiological disorders
like hypertension and psychological illness such as anxiety disorder.
Psychological stress is a growing concern for the worldwide population across
all age groups. A reliable, cost efficient, acute stress detection system could
enable its users to better monitor and manage their stress to mitigate its long-
term negative effects. In this article, we will review and discuss the literature
that has used machine learning based approaches for stress detection. We will
also review the existing solutions in the literature that have leveraged the
concept of edge computing in providing a potential solution in real-time
monitoring of stress.
Introduction
Stress is one of the main factors that are affecting millions of lives. Mental
stress needs to be in control as it results in different dangerous suffering. Timely
mental stress detection can help to prevent stress related health problems. It is
important to inform the person about his/her unhealthy life style and even alarm
him/her before any acute condition occurs. Stress detection technology could
help people better understand and relieve stress by increasing their awareness of
heightened level of stress that would otherwise go undetected. For this objective,
we design a smart device that has the capability of monitoring stress level
continuously in everyday activities. It not only sends alert to users via
smartphone but allows them to manage stress and share health data with doctors
and family. Remote patient monitoring using this device may increase access to
efficient caring and decrease health care delivery cost.

LITERATURE SURVEY

1.Title: New Avenues in Opinion Mining and Sentiment Analysis

Author: E. Cambria, B. Schuller, Y. Xia, and C. Havasi
Abstract: Others’ opinions can be crucial when it’s time to make a decision or
choose among multiple options. When those choices involve valuable resources
(for example, spending time and money to buy products or services) people
often rely on their peers’ past experiences. Until recently, the main sources of
information were friends and specialized magazine or websites. Now, the “social
web” provides new tools to efficiently create and share ideas with everyone
connected to the World Wide Web. Forums, blogs, social networks, and content-
sharing services help people share useful information. This information is
unstructured, however, and because it’s produced for human consumption, it’s
not something that’s “machine process able.” Capturing public opinion about
social events, political movements, company strategies, marketing campaigns,
and product preferences is garnering increasing interest from the scientific
community (for the exciting open challenges), and from the business world (for
the remarkable marketing fallouts and for possible financial market prediction).
The resulting emerging fields are opinion mining and sentiment analysis.
Disadvantages:
In existing opinion-mining systems need to be improve to broader and deeper
common and commonsense knowledge bases. More complete knowledge must
be combined with reasoning methods that are more deeply inspired by human
thought and psychology. This will lead to a better understanding of natural
language opinions and will more efficiently bridge the gap between
(unstructured) multimodal information and (structured) machine-process able
data.

2. Title: New avenues in knowledge bases for natural language processing

Author: E. Cambria, B. Schuller, Y. Xia, and and B. White
Abstract: Between the birth of the Internet and 2003, year of birth of social
networks such as Myspace, Delicious, LinkedIn, and Facebook, there were just a
few dozen Exabyte’s of information on the Web. Today, that same amount of
information is created weekly. The advent of the Social Web has provided
people with new content sharing services that allow them to create and share
their own contents, ideas, and opinions, in a time- and cost-efficient way, with
virtually millions of other people connected to the World Wide Web. This huge
amount of information, however, is mainly unstructured (because it is
specifically produced for human consumption) and hence not directly machine-
process able. The automatic analysis of text involves a deep understanding of
natural language by machines, a reality from which we are still very far off.
Hitherto, online information retrieval, aggregation, and processing have mainly
been based on algorithms relying on the textual representation of web pages.
Such algorithms are very good at retrieving texts, splitting them into parts,
checking the spelling and counting the number of words. When it comes to
interpreting sentences and extracting meaningful information, however, their
capabilities are known to be very limited, as most of the existing approaches are
still based on the syntactic representation of text, a method that relies mainly on
word co-occurrence frequencies. Such algorithms are limited by the fact that
they can process only the information that they can ‘see’. As human text
processors, we do not have such limitations as every word we see activates a
cascade of semantically related concepts, relevant episodes, and sensory
experiences, all of which enable the completion of complex natural language
processing (NLP) tasks – such as word-sense disambiguation, textual entailment,
and semantic role labeling – in a quick and effortless way.
Disadvantages:
This doesn't take context and content information for analyzing sentiments.
Accuracy is too low.

3. Title: Affective Computing and Sentiment Analysis

Author: E. Cambria
Abstract: Emotions play an important role in successful and effective human–
human communication. In fact, in many situations, emotional intelligence is
more important than IQ for successful interaction. There is also significant
evidence that rational learning in humans is dependent on emotions.2 Affective
computing and sentiment analysis, hence, are key for the advancement of AI3
and all the research fields that stem from it. Moreover, they find applications in
various scenarios and companies, large and small that include the analysis of
emotions and sentiments as part of their mission. Sentiment-mining techniques
can be exploited for the creation and automated upkeep of review and opinion
aggregation websites, in which opinionated text and videos are continuously
gathered from the Web and not restricted to just product reviews, but also to
wider topics such as political issues and brand perception.
Disadvantages:
So far, sentiment-mining approaches from text or speech have been based
mainly on the bag-of-words model because, at first glance, the most basic unit of
linguistic structure appears to be the word.

4. Title: A time-varying propagation model of hot topic on BBS sites and

Blog networks
Author: B. Zhang, X. Guan, M. J. Khan, and Y. Zhou
Abstract: Modeling the propagation of hot online topic is a preliminary
requirement of predicting the trend of hot online topic. We propose a time-
varying hot topic propagation model in online discussion context based upon the
collective behavior of users who are in different social subgroups on blog
networks and bulletin board system (BBS) sites. By analyzing the stability of the
equilibrium of our model, we search for the threshold to be watershed of the
trend of hot online topic and generalize about two theorems from the results of
analysis, they exposit two sufficient conditions under which the trend of hot
online topic will die out or remain uniformly weakly persistent. Furthermore, we
propose methods to predict the trend of hot online topic on the strength of our
model and theorems. For different motivation, we design two methods: Method
(I) is mainly served as a way of theoretical research for predicting long trend of
single-peak hot online topic by the thresholds of theorems; and for application,
we design method (II) to predict the number of users writing or commenting
upon article posts with respect to multi-peak hot online topic and single-peak
one in the following two days with the help of Method (I). Experiments of two
methods are performed on widely-discussed topics on the Sina Blog and the
famous Liang Quan Qi Mei (LQQM) BBS and Xi'an Jiao tong University
(BMY) BBS in China. The experimental results show that our methods predict
the trend of hot online topic efficiently not only for theoretical motivation but
also for applicable motivation, and reduce the computational complexity. Hence,
our model can serve as basis for predicting trends in hot online topic
propagation.
Disadvantages: There is no Sentiment analysis done in social network.
OBJECTIVES:
1. To build an efficient IOT and ML based system for accurate classification of
stress.
2. To identify the stress of the user based on various parameters.
3. To alert or notify the user if there is variation in the stress and asking them to
take the necessary precautions.
4. To notify the user’s family and his/her doctors in case of major changes or if
there is chance for an heart attack to occur.

METHODOLOGY

In this project, the data is collected using IOT sensors such as temperature
sensor, skin conductance sensor and accelerometer. Skin conductance sensor is
used as physical activities stimulate sweat glands which can bring variation in
skin conductance. There is a data collection phase to collect the training
examples needed for the Logical regression algorithm. The collected data is
then fed into the appropriate Logical regression algorithm, which classifies
these set of parameters along with the user, into one of the stress levels. If any
abnormality in the stress level is detected, the user is notified on his smart phone
text message and is asked to take the necessary precautions. If there are major
changes in the stress level that can cause harm to the user’s health, or if any
chance of heart attack is predicted, the user, his emergency contacts and his
doctor are notified.
EXISTING SYSTEM

 The existing system for detection of stress is Perceived Stress Scale(PSS).

 It is a measure of the degree to which situations in ones life are appraised
as stressful.
 The scale also includes a number of direct queries about current levels of
experienced stress.

PROPOSED SYSTEM

 The proposed system will attempt to create a device that monitors stress in
a continuous manner where the heart beat rate (using pulse sensor) and
skin conductance are used to determine the stress level.
 The accelerometer sensor is a device that is used to measure acceleration
forces. Such forces maybe static or dynamic to sense movements or
vibrations.
 These sensor information are collectively stored in the cloud.
 Now, using an appropriate ML algorithms (like Logistic Regression), we
do the prediction on the data that is stored and the predicted results falls
into various classes such as undertrained, over trained etc.
BLOCK DIAGRAM

SMPS 12V DC LM317 5V DC

230V AC
ADAPTOR SUPPLY SUPPLY

LCD

Accelerometer RENESAS Skin conductance

MICROCONTROLLER
Smoke sensor Buzzer

PYTHON Temperature
LAPTOP
(LM35)

HARDWARE REQUIREMENTS:
 Computer – Processor, high speed is preferred. 64 - bit
 Hard disk – Free space of 5GB
 Laptop built-in Camera
 Renesas Microcontroller
 LCD display for Microcontroller
 Temperature sensor(LM35)
 Accelerometer
 Smoke Sensor
 Skin Conductance Sensor
 Buzzer
SOFTWARE’S USED:

1. Python 3.7
2. Open CV tool
3. OS – Windows 8 / 10, 64 – bit.
4. Embedded C
5. Cube Suite+ Compiler
6. Renesas Flash Programmer V2.0

STRESS DATASET:

Attributes in the dataset

Userid  dataset collected from multiple users.

Acceleration  0 – user Normal condition, 1 – user in active condition like

jogging, steps climbing, doing more physical work.

Temperature 30’C above is normal, 25’C below is stress

Heartrate  74 below in normal, 75 above is stress

Skin conductance  100 above in normal, 200 above is stress

Label  0- Normal, 1 – stress.

FUTURE WORK
Most of the works on stress detection are oriented towards psychological
stress detection. Although there have been few works on physiological stress
detection, there is further need for research towards detecting physiological
stress. Research should be conducted to design robust classification models that
can generalize classifications irrespective of the signal acquisition device and
configuration. This will ensure the applicability of a framework in any setting
irrespective of the type of edge devices and the configuration in which the data is
collected.
Another interesting research direction that could be explored is the use of
an edge-cloud framework which will distribute the process of pre-processing,
feature extraction and classification across different layers of edge and cloud.
This might be useful when integrating multiple end users to a single platform. In
this context, there are other research scopes such as dealing with the massive
amount of data that might be generated at the edge with time. Works are being
done in this field to enhance the quality of data storing and retrieval mechanisms
in an edge cloud settings. Ensuring privacy, security, and confidentiality of the
data streams generated is also another exciting research direction. Another
direction that would be interesting to explore is to study the effectiveness of
different intervention mechanisms in mediating acute stress responses. This will
be a critical step towards developing a ubiquitous stress detection and
management system which has the potential to further develop into a smart
health monitoring system. Finally, real-time monitoring of stress can be
integrated in a smart home environment to assist older adults and persons with
dementia or cognitive impairment.
RESULTS AND CONCLUSIONS
1. A dataset containing the stress classifications for a set of parameters.
2. Predicting the stress levels using ML algorithms with good accuracy.
3. A system that monitors changes in stress levels of a user in their everyday
activities and notifying them on these changes.

REFERENCES
[1] S. Cohen, R. C. Kessler, and L. U. Gordon, Strategies for measuring stress in
studies of psychiatric and physical disorders. Oxford University Press, 1995.
[2] B. Mcewen, “Protective and damaging effects of stress mediators: Central
role of the brain,” Dialogues in clinical neuroscience, vol. 8, no. 02, pp. 367–81,
2006.
[3] P. Adams, M. Rabbi, T. Rahman, M. Matthews, A. Voida, G. Gay, T.
Choudhury, and S. Voida, “Towards personal stress informatics: Comparing
minimally invasive techniques for measuring daily stress in the wild,” in
Proceedings of the 8th International Conference on Pervasive Computing
Technologies for Healthcare, 2014, pp. 72– 79.
[4] S. Greene, H. Thapliyal, and A. Caban-Holt, “A survey of affective
computing for stress detection: Evaluating technologies in stress detection for
better health,” IEEE Consumer Electronics Magazine, vol. 5, no. 4, pp. 44–56,
Oct 2016.
[5] H. Thapliyal, V. Khalus, and C. Labrado, “Stress detection and management:
A survey of wearable smart health devices,” IEEE Consumer Electronics
Magazine, vol. 6, no. 4, pp. 64–69, Oct 2017.
[6] J. Bakker, M. Pechenizkiy, and N. Sidorova, “What’s Your Current Stress
Level? Detection of Stress Patterns from GSR Sensor Data,” in Proceedings
IEEE 11th International Conference on Data Mining Workshops, 2011, pp. 573–
580.
[7] A. de Santos Sierra, C. Sanchez Avila, J. Casanova, and G. Bailador, “A
Stress-Detection System Based on Physiological Signals and Fuzzy Logic,”
IEEE Transactions on Industrial Electronics, vol. 58, pp. 4857–4865, November
2011.
[8] C. Setz, B. Arnrich, J. Schumm, R. La Marca, G. Trà ˝uster, and U. Ehlert,
“Discriminating Stress From Cognitive Load Using a Wearable EDA Device,”
IEEE Transactions on Information Technology in Biomedicine, vol. 14, no. 2,
pp. 410–417, 2010.
[9] S. Yoon, J. Kyoung Sim, and Y.-H. Cho, “A flexible and wearable human
stress monitoring patch,” Scientific Reports, vol. 6, p. 23468, 03 2016.