Assessing Accuracy in Diagnosing Narcissistic Personality Disorder

with Elastic Net Classifier versus Lasso Regression Classifier

Navali Indravathi and A. Jegatheesan

Saveetha University, India

Keywords: Diagnosis, Health, Lasso Regression Classifier, Machine Learning, Narcissistic Personality Disorder, Novel

Elastic Net Classifier, Ridge Regression.

Abstract: The objective of the study is to compare the Lasso Regression Classifier with an Elastic Net Classifier to

identify narcissistic personality disorder. Materials and Methods: With a limit of 0.05% and datasets collected

from various web sources with continuing reviews, a novel elastic net method and lasso regression technique

were used with g-power value 80%. Data was obtained after 10 iterations of the confidence span 95%. Results:

Elastic Net Algorithm prediction accuracy is 86%, and Lasso Regression prediction accuracy is 67%. The

prediction difference is statistically significant at p = 0.001, (p<0.05). This suggests that the two algorithms

differ statistically significantly from one another. Conclusion: The conclusion is Novel Elastic Net Classifier

performs more accurately than the Lasso Regression Classifier.

1 INTRODUCTION

A sign of narcissistic personality disorder is a

propensity to exaggerate or boast about successes in

an effort to impress others. This tendency is linked to

a need for a lot of praise and attention from other

people (Fan et al. 2021). By providing expert

diagnosis, health treatment, and guidance, the

material on narcissistic personality disorder aims to

help people prevent self-harm or mental health crises

(Ramos-Lima et al. 2022). People may learn about

self harm using these applications, which also

encourage a life dedicated to preventing suicide

(Howard 2013). Numerous health care professionals

find it beneficial to comprehend the severe conditions

in which someone is experiencing narcissism (Eken

et al. 2022) (G. Ramkuamr et al. 2021).

The previous five years have seen a total of 78

articles on sciencedirect and 65 on researchgate.net

written on this topic. This study's main objective is to

perform a meta-analysis to discriminate between

those with bipolar illness and health controls

(Oliveira et al. 2019). The multimodal temporal

machine learning algorithms for borderline disorder

are designed to recognise the depressed state of an

individual's health and interpret behavioral patterns in

people (Xiong et al. 2020). Finding their non-

affective psychiatric symptoms is the largest obstacle

a person faces when forecasting the abnormality of

their health and concentrating on those components

of their memory (Sawalha et al. 2021) (James et al.

2017). Combining a natural language method with

machine learning classifiers to analyze mood

dynamics allows for the detection of mood swings,

self-centeredness, and a tendency to ignore the

opinions or feelings of others (Vrabie et al. 2013)

(Sivakumar et al 2022). Effective brain connections

allow for the treatment of narcissistic disorder

through an accurate technique of diagnosis and offer

great promise as a therapeutic tool (Garland, Jane

Garland, and Duffy 2010).

There are several problems with the existing

approach since each person's response will be

different given that the research is about individuals'

health with narcissistic personality disorder. With the

use of the right questionnaire session, narcissistic

personality disorder will be more accurately

diagnosed, according to this study's goal. This will be

accomplished by using an unique elastic net classifier

in addition to a lasso regression classifier to get more

accurate results.

Indravathi, N. and Jegatheesan, A.

Assessing Accuracy in Diagnosing Narcissistic Personality Disorder with Elastic Net Classiﬁer versus Lasso Regression Classiﬁer.

DOI: 10.5220/0012559900003739

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

In Proceedings of the 1st International Conference on Artiﬁcial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics (AI4IoT 2023), pages 77-82

ISBN: 978-989-758-661-3

2 MATERIALS AND METHODS

This research evaluation was carried out in the

Saveetha Institute of Medical and Technical Sciences

(SIMATS), Saveetha School of Engineering of

machine learning lab. In this article there are two

groups that are defined. A Novel Elastic Net classifier

for the first group and a lasso regression classifier for

the second group are provided. Using previous

evaluations from clinical.com, the test size for each

component will be set using a constant g power of

80%, a boundary of 0.05, and an inevitability of 95%.

The dataset information was obtained from recent

research work of kaggle (Malik and Mumtaz 2019).

The example size of the Novel Elastic Net model

(N=10) and the Ridge regression (N=10) will both be

intentionally chosen (Kane, Phar, and BCPS).

2.1 Elastic Net Classifier

The innovative elastic net classifier, a machine

learning algorithm, selects a regression analysis

utilizing both the lasso regression and the ridge

regression classifiers. A unique elastic net classifier

is used to predict the data (such as name, age, and

score) and compute the scores of the data using lasso

regression and ridge regression in order to identify

whether a person has narcissistic personality disorder.

To comprehend the ground-breaking elastic net

classifier, one must have a firm knowledge of lasso

regression and ridge regression techniques. By

efficiently deleting any unnecessary information,

lasso regression and the lasso regression classifier

simplify the data. A value of 0 for lasso regression

and a value of 1 for ridge regression are equivalent.

The data is predicted by the help of ridge regression

and lasso regression algorithm.

Pseudo Code: Elastic Net Classifier

1. The data from training are X and Y.

2. Data is in the (0–40) range.

3. Build a website application with 40

attributes ranging from 1 to 40

4. Each attribute has the first and second

phrases (1 and 2).

5. Consistently provide the attribute

declarations values.

6. For each attribute value, enter a (1 or 2)

statement.

7. Last but not least, insert the gender and age

in the characteristics field.

8. Figuring out the scores of attributes.

Pseudo Code: Lass Regression Classifier

1. The information from training are X and Y

2. Data is in the range of 0 and 40.

3. Create a website application that has 40

qualities, ranging from 1 to 40.

4. Each characteristic includes both the first

and second statements (1 and 2).

5. Regularly provide attribute declarations

values.

6. For each attribute value, provide a

statement of type (1 or 2).

7. Enter the gender and age in the attributes

section.

8. Obtain the scores of attributes.

2.2 Lasso Regression Classifier

Table 1: Raw data table for evaluating the accuracy of

Elastic Net Classifier and Lasso Regression Classifier.

S.No

Elastic Net Classifier

Ridge Regression

Classifier

86.00

67.00

85.50

66.50

84.00

66.10

83.40

66.00

82.60

65.70

81.50

65.00

80.03

64.70

80.00

63.30

79.80

62.90

79.40

61.90

Table 2: Group statistics results (mean of novel elastic net

classifier 82.22 is greater than the lasso regression classifier

664.91 and standard. Error Mean for ENC is 0.77319 and

LRC is 0.53613).

Groups

Mean

Std. Deviation

Std. Mean

Error

Accuracy

82.2230

2.44503

.77319

KNN

64.9100

1.69539

.53613

AI4IoT 2023 - First International Conference on Artiﬁcial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry

and Consumer Electronics

Table 3: Independent Sample T-test Result is done with confidence interval as 95% and significance level as 0.001 (novel

elastic net classifier appears to perform significantly better than lasso regression classifier with the value of p<0.05).

Independent Sample Test

Levene’s Test for Equality of Variances

T-test for Equality of Means

Sig.

(2-

tailed)

Mean

Difference

Std. Error

Differences

95% Confidence

Interval of the

Difference

Lower

Upper

Accuracy

Equal Variances assumed

2.539

.008

18.4

.001

17.31

0.940

15.33

19.28

Equal Variances not

assumed

18.4

16.02

.001

17.31

0.940

15.313

9.307

Figure 1: Novel Elastic net classifier and lasso regression classifier in terms of mean accuracy. Mean accuracy of nove elastic

net algorithm (86%) is better than lasso regression classifier (67%). Standard deviation of novel elastic net algorithm is

slightly better than lasso regression X Axis: ENC vs LRC. Y Axis: Mean accuracy of detection = +/- 2 SD with a Confidence

Interval of 95%.

Lasso regression classifier may be used as a statistical

method to find factors that are predictive of

narcissistic personality disorder and to comprehend

the connection between these factors and the result.

For instance, lasso regression classifier may be used

in a study to pinpoint clinical and societal variables

linked to the emergence of narcissistic personality

disorder and to ascertain how these variables affect

the condition's trajectory. Researchers may be better

able to comprehend the underlying causes of the

condition and create more specific and practical

remedies with the use of lasso regression classifier in

this situation. The use of a single statistical technique,

such as the lasso regression classifier is not

recommended since narcissistic personality disorder

is a complex condition with various origins.

Apply the generated approach and the elastic net

technique to the training dataset. By combining all the

data, the errors and incorrect inputs were eliminated.

At this moment, the preprocessing was done. After

that, the data set is split into two groups: one for

training and one for testing. 25% more system

training is included in this dataset, which comprises

75% testing. In order to improve and reach high

accuracy, test sets are produced after analyzing the

computations carried out throughout two interaction

trains. The data set concerns the identification of

narcissistic personality disorder. 11,000 details make

up the dataset. The information set is split into 25%

Assessing Accuracy in Diagnosing Narcissistic Personality Disorder with Elastic Net Classiﬁer versus Lasso Regression Classiﬁer

for training and 75% for testing. Python-based

machine learning falls under the category of

narcissistic personality disorder. The dataset

requirements to run this. Python and the flask library's

web frame are used in the software specification for

narcissistic personality disorder. The use of machine

learning technologies by the Windows operating

system and the Jupyter notebook's IDE.

The examination was conducted using IBM SPSS

version 21. It is a tool for the quantitative creation of

software that analyzes data. Ten recursions were

completed with the help of a limit of 10–20 models

and two current and recommended cunnings. To

evaluate the accuracy for each recursion was

gathered, the information was gleaned from recent

Kaggle research. A processor with a base clock of 3.8

GHz or greater, 4 GB of RAM, and a minimum of 800

MB of disc space are the minimum hardware.

3 STATISTICAL ANALYSIS

The independent variables for this test are the test

name, step and language type which remain

consistent even when the dataset size is increased and

the limits and components are changed in accordance

with the data sources and the components are moved

to account for any changes in the data. For the

purpose of diagnosing narcissistic personality

disorder, the autonomous T-test is utilized to contrast

the lasso regression approach to the elastic net

approach. In contrast to age and score are independent

variables and accuracy is a dependent variable.

4 RESULTS & DISCUSSION

Elastic Net Classifier and Lasso Regression Classifier

accuracy was compared in Table 1 using raw data.

The error of the innovative elastic net classifier and

the lasso regression classifier based on the diagnosis

of narcissistic personality disorder are shown in Table

2 respectively. It demonstrates that the lasso

regression classifier calculation will have an

exactness mean of 64.91%, Standard Deviation 1.69,

while the revolutionary elastic net classifier technique

will have a mean of 82.22%, Standard Deviation 2.44.

The mean and standard deviation of the lasso

regression classifier and the novel elastic net

classifier of narcissistic personality disorder are

tabulated in Table 3, which highlights the substantial

difference between the two groups with

p=0.000(p<0.05).

The mean accuracy of a novel elastic net classifier

and lasso regression classifiers based on the diagnosis

of narcissistic personality disorder are contrasted in

Fig. 1.

The experiments revealed that the novel elastic

net classifier outperformed the lasso regression

strategy. The accuracy of the lasso regression

classifier and the novel elastic net classifier is

calculated using the SPSS tool. The lasso regression

technique with 86% and the original, creative elastic

net classifier with 67% has the accuracy to diagnosis

of narcissistic personality disorder in the dataset. The

statistical analysis calculated in narcissistic

personality disorder difference p=0.001 (p<0.05)

tested by using t-test states values of results are

significant.

The prediction, reduce treatment costs, and

prevent life-threatening illnesses, supporting research

is conducted on narcissistic personality disorder.

(Sonkurt et al. 2021). The identification of the

condition includes a few recognition computations

required to comprehend narcissistic personality

disorder (Paris 2017). To counteract this research,

using machine learning techniques to raise the

accuracy by expanding the features of the medical

dataset(Paris 2017; Colombo et al. 2022). The result

and precision of the information in the example given

form the basis for this capability to recognize items

(Ravan et al. 2023). Introduce a straightforward

classifier to evaluate and better understand

personality disorders (Parker et al. 2022).With the use

of vast datasets and high accuracy classifiers like new

elastic, this classifier successfully identifies the

disease (Walsh-Messinger et al. 2019). This project

will use machine learning methods to investigate

narcissistic disease. (Mumtaz et al. 2015).

Accuracy will be established by constraints like

certain scores and the volume of data. The accuracy

percentages will be determined by using the term

"narcissistic personality disorder" as well as a

collection of data that includes the size and score of

each condition. If the average mistake could be greatly

decreased, that would be desirable. It is required to

optimize the data related to this recognition. As a

consequence, the work performance offered by the

innovative elastic net classifier has more accuracy and

a smaller mean error when compared to the lasso

regression technique. To maximize data, this method

will be used in the future to improve the diagnostic and

classification accuracy of narcissistic personality

disorder. future algorithms for high-quality selection.

In future, quality choice algorithms may be utilized to

decrease the calculation time and develop the

diagnosing accuracy of classifiers.

AI4IoT 2023 - First International Conference on Artiﬁcial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry

and Consumer Electronics

5 CONCLUSIONS

When combined with a lasso regression classifier

which offers (67%) accuracy, the innovative elastic

net classifier delivers greater accuracy (86%) in light

of collected results. This recognition associated data

needs to be optimized. Samples of the data should be

gathered from continuous sites like Kaggle. The

narcissistic personality disorder diagnosis will be

used to determine the accuracy values and it also

counts the size and score of each narcissistic

personality disorder in a set of data. It would be

preferable if the average error could be significantly

reduced.

REFERENCES

Colombo, Federica, Federico Calesella, Mario Gennaro

Mazza, Elisa Maria Teresa Melloni, Marco J. Morelli,

Giulia Maria Scotti, Francesco Benedetti, Irene

Bollettini, and Benedetta Vai. 2022. “Machine

Learning Approaches for Prediction of Bipolar

Disorder Based on Biological, Clinical and

Neuropsychological Markers: A Systematic Review

and Meta-Analysis.” Neuroscience & Biobehavioral

Reviews. https://doi.org/10.1016/j.neubiorev.

2022.104552.

Eken, Aykut, Damla Sayar Akaslan, Bora Baskak, and

Kerim Münir. 2022. “Diagnostic Classification of

Schizophrenia and Bipolar Disorder by Using Dynamic

Functional Connectivity: An fNIRS Study.” Journal of

Neuroscience Methods. https://doi.org/10.1016/j.

jneumeth.2022.109596.

Fan, Ru, Tiantian Hua, Tian Shen, Zhigang Jiao, Qingqing

Yue, Bingwei Chen, and Zhi Xu. 2021. “Identifying

Patients with Major Depres sive Disorder Based on

Tryptophan Hydroxylase-2 Methylation Using

Machine Learning Algorithms.” Psychiatry Research.

https://doi.org/10.1016/j.psychres.2021.114258.

Garland, E. Jane, E. Jane Garland, and Anne Duffy. 2010.

“Treating Bipolar Disorder in the Early Stages of

Illness.” Practical Management of Bipolar Disorder.

https://doi.org/10.1017/cbo9780511776922.008.

G. Ramkumar, R. Thandaiah Prabu, Ngangbam Phalguni

Singh, U. Maheswaran, Experimental analysis of brain

tumor detection system using Machine learning

approach, Materials Today: Proceedings, 2021, ISSN

2214-7853, https://doi.org/10.1016/j.matpr.2021.01.

246

Howard, Newton. 2013. “Approach Towards a Natural

Language Analysis for Diagnosing Mood Disorders

and Comorbid Conditions.” 2013 12th Mexican

International Conference on Artificial Intelligence.

https://doi.org/10.1109/micai.2013.50.

James, J., Lakshmi, S. V., & Pandian, P. K. (2017). A

preliminary investigation on the geotechnical properties

of blended solid wastes as synthetic fill material.

International Journal of Technology, 8(3), 466-476.

Kane, Sean P., Phar, and BCPS. n.d. “Sample Size

Calculator.” Accessed April 19, 2023.

https://clincalc.com/stats/samplesize.aspx.

Malik, Aamir Saeed, and Wajid Mumtaz. 2019.

“Introduction: Depression and Challenges.” EEG-

Based Experiment Design for Major Depressive

Disorder. https://doi.org/10.1016/b978-0-12-817420-

3.00001-1.

Mumtaz, Wajid, Aamir Saeed Malik, Syed Saad Azhar Ali,

and Mohd Azhar Mohd Yasin. 2015. “P300 Intensities

and Latencies for Major Depressive Disorder

Detection.” 2015 IEEE International Conference on

Signal and Image Processing Applications (ICSIPA).

https://doi.org/10.1109/icsipa.2015.7412250.

Oliveira, Leticia de, Leticia de Oliveira, Liana C. L.

Portugal, Mirtes Pereira, Henry W. Chase, Michele

Bertocci, Richelle Stiffler, et al. 2019. “Predicting

Bipolar Disorder Risk Factors in Distressed Young

Adults From Patterns of Brain Activation to Reward: A

Machine Learning Approach.” Biological Psychiatry:

Cognitive Neuroscience and Neuroimaging.

https://doi.org/10.1016/j.bpsc.2019.04.005.

Paris, Joel. 2017. “Differential Diagnosis of Bipolar

Disorder and Borderline Personality Disorder.” Bipolar

Disorders. https://doi.org/10.1111/bdi.12565.

Parker, Gordon, Michael J. Spoelma, Gabriela Tavella,

Martin Alda, David L. Dunner, Claire O’Donovan,

Janusz K. Rybakowski, et al. 2022. “A New Machine

Learning-Derived Screening Measure for

Differentiating Bipolar from Unipolar Mood

Disorders.” Journal of Affective Disorders.

https://doi.org/10.1016/j.jad.2021.12.070.

Ramos-Lima, Luis Francisco, Vitoria Waikamp, Thauana

Oliveira-Watanabe, Mariana Recamonde-Mendoza,

Stefania Pigatto Teche, Marcelo Feijo Mello, Andrea

Feijo Mello, and Lucia Helena Machado Freitas. 2022.

“Identifying Posttraumatic Stress Disorder Staging

from Clinical and Sociodemographic Features: A

Proof-of-Concept Study Using a Machine Learning

Approach.” Psychiatry Research.

https://doi.org/10.1016/j.psychres.2022.114489.

Ravan, M., A. Noroozi, M. Margarette Sanchez, L. Borden,

N. Alam, P. Flor-Henry, and G. Hasey. 2023.

“Discriminating between Bipolar and Major Depressive

Disorder Using a Machine Learning Approach and

Resting-State EEG Data.” Clinical Neurophysiology.

https://doi.org/10.1016/j.clinph.2022.11.014.

Sawalha, Jeffrey, Liping Cao, Jianshan Chen, Alessandro

Selvitella, Yang Liu, Chanjuan Yang, Xuan Li, et al.

2021. “Individualized Identification of First-Episode

Bipolar Disorder Using Machine Learning and

Cognitive Tests.” Journal of Affective Disorders.

https://doi.org/10.1016/j.jad.2020.12.046.

Sivakumar, V. L., Nallanathel, M., Ramalakshmi, M., &

Golla, V. (2022). Optimal route selection for the

transmission of natural gas through pipelines in

Tiruchengode Taluk using GIS–a preliminary study.

Materials Today: Proceedings, 50, 576-581.

Assessing Accuracy in Diagnosing Narcissistic Personality Disorder with Elastic Net Classiﬁer versus Lasso Regression Classiﬁer

Sonkurt, Harun Olcay, Ali Ercan Altınöz, Emre Çimen,

Ferdi Köşger, and Gürkan Öztürk. 2021. “The Role of

Cognitive Functions in the Diagnosis of Bipolar

Disorder: A Machine Learning Model.” International

Journal of Medical Informatics.

https://doi.org/10.1016/j.ijmedinf.2020.104311.

Vrabie, M. I., I. Miclutia, V. Marinescu, and C. Popescu.

2013. “1983 – Identifying Specific Domains of

Cognitive Dysfunction for Manic Episodes in Bipolar

Affective Disorder.” European Psychiatry.

https://doi.org/10.1016/s0924-9338(13)76920-7.

Walsh-Messinger, Julie, Haoran Jiang, Hyejoo Lee, Karen

Rothman, Hongshik Ahn, and Dolores Malaspina.

2019. “Relative Importance of Symptoms, Cognition,

and Other Multilevel Variables for Psychiatric Disease

Classifications by Machine Learning.” Psychiatry

Research. https://doi.org/10.1016/j.psychres.2019.03.

048.

Xiong, Zihan, Xuanhua Zhang, Mengxuan Zhang, and

Bosen Cao. 2020. “Predicting Features of Human

Mental Disorders through Methylation Profile and

Machine Learning Models.” 2020 2nd International

Conference on Machine Learning, Big Data and

Business Intelligence (MLBDBI).

https://doi.org/10.1109/mlbdbi51377.2020.00019.

AI4IoT 2023 - First International Conference on Artiﬁcial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry

and Consumer Electronics