|Year : 2022 | Volume
| Issue : 1 | Page : 28-34
Multilayer Perceptron to Assess the Impact of Anatomical Risk Factors on Traumatic Dental Injuries: An Advanced Statistical Approach of Artificial Intelligence in Dental Traumatology
Mohammad Kamran Khan BDS, MDS 1, Mahendra Kumar Jindal2
1 Private Pediatric Dental Practice, Aligarh, Uttar Pradesh, India
2 Department of Pediatric and Preventive Dentistry, Dr. Ziauddin Ahmad Dental College and Hospital, Aligarh Muslim University, Aligarh, Uttar Pradesh, India
|Date of Submission||06-Feb-2021|
|Date of Decision||13-May-2022|
|Date of Acceptance||23-May-2022|
|Date of Web Publication||05-Aug-2022|
Dr. Mohammad Kamran Khan
Specialist Consultant Pedodontist, Hamdard Nagar-A, Civil Lines, Aligarh, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Introduction: Traumatic dental injuries (TDIs) are the public dental health concern, with variable prevalence reported worldwide. Although, TDI is not a disease rather, it is a result of various risk factors. This study was performed to assess the influence of anatomical risk factors such as accentuated overjet, overbite, molar relationship, and lip competency in determining the number of traumatized teeth per affected individual by using the advanced statistical method of multilayer perceptron (MLP) model of deep learning algorithm of artificial intelligence (AI). Materials and Methods: A cross-sectional study consisted of 1000 school children (boys and girls) of index age groups between 12 and 15 years selected through multistage sampling technique. Orofacial anatomical risk factors associated with TDI were statistically analyzed by MLP model of deep learning algorithm of AI using IBM SPSS Modeler software (version 18, 2020). Results: MLP method revealed results in terms of normalized importance as overbite (100%) was the strongest risk factor for the occurrence of TDI in number of teeth of affected participants, followed by molar relationship (90.2%), overjet (87.7%), and the lip competency was found as the weakest risk factor. Conclusion: Using the MLP as statistical method, overbite was found as the strongest anatomical risk factor in determining the number of traumatized teeth per affected individual as compared to molar relationship, overjet, and lip competence.
Keywords: Advanced statistical analysis, anatomical risk factors, artificial intelligence (AI), artificial neural network (ANN), cross-sectional study, deep learning, machine learning, modern statistical methods, multilayer perceptron (MLP), traumatic dental injuries
|How to cite this article:|
Khan MK, Jindal MK. Multilayer Perceptron to Assess the Impact of Anatomical Risk Factors on Traumatic Dental Injuries: An Advanced Statistical Approach of Artificial Intelligence in Dental Traumatology. J Orofac Sci 2022;14:28-34
|How to cite this URL:|
Khan MK, Jindal MK. Multilayer Perceptron to Assess the Impact of Anatomical Risk Factors on Traumatic Dental Injuries: An Advanced Statistical Approach of Artificial Intelligence in Dental Traumatology. J Orofac Sci [serial online] 2022 [cited 2022 Aug 7];14:28-34. Available from: https://www.jofs.in/text.asp?2022/14/1/28/353474
| Introduction|| |
Oral health is also an integral part of the individual’s overall health and hence, any disease or trauma to oral and dental tissues influences the overall health of the individual. Traumatic dental injuries (TDIs) are the public dental health concern, with variable prevalence reported worldwide (6%–59%). Earlier, it was assumed that dental injuries would most likely exceed dental caries and periodontal diseases and now that seem to be true by seeing their differences in prevalence in recent studies. TDIs are the most commonly affected injuries in children and adolescents, making them highly susceptible age group for dental trauma. It has been observed that such dental injuries are more common in certain age groups, but no individual can ever be at zero risk through their daily living activities.
Although, TDI is not a disease, it is a result of various risk factors related to an individual’s life. The consequences of such injuries not only affect an individual physically or economically but also increases the psychosocial burden indefinably. TDIs are costly and time consuming to manage as compared to other bodily injuries in emergency hospitals. The average number of visits needed to treat such injuries are 1.9 to 9.1 while for other bodily traumas are just 1.5.
The approach for managing dental caries and oral diseases has evolved from “treating the lesions” to “managing risk factors” associated with them. Likewise, the risk factors related to TDIs need to be identified comprehensively by all the possible ways to halt such injuries to occur. So far, several risk factors for dental trauma have been reported in literature such as age, gender, geographical location, lip incompetency, and accentuated overjet, anteroposterior molar relation, overbite, ethnic group, socioeconomic status, obesity, peer relationships, family type, school grade, physical activity level, perception of paternal punishment, birth order, and psychosocial factors., These can be grouped into oral anatomical factors, environmental factors, and behavioral factors, etc.
Among orofacial anatomical risk factors, inadequate lip coverage and accentuated overjet have been studied extensively as a predictor for TDI, while overbite and molar relation have not been studied to that extent. Very few studies have investigated the association and strength of effect of all anatomical risk factors affecting TDI, and in most of them regression analysis method has been applied.
The multivariable analysis methods (i.e., multiple linear or logistic regressions statistics) have been used much for studying the associations of the variables. But, the limitations of them are the explanatory variables (covariates) occurrences when the clinical data are analyzed. However, the concerns regarding the collinearity and multicollinearity which bring about forged results obtained due to multivariable statistical analysis have been overlooked in the dental research. Science is all about for exploration of innovative tools.
Newer analysis techniques should be used to investigate the complex aspects of risk factors of TDI. One such newer analysis method based on deep learning of artificial intelligence (AI) is the artificial neural network (ANN) that can be employed to perform nonlinear statistical modeling as new alternative to logistic regression.
Initially, AI was supposed to be related to only robots or computers, but its applications are also found or evolved in the fields of medicine, dentistry, philosophy, linguistics, psychology, and statistics.- Various applications of AI in medical field have been reported such as: early detection of atrial fibrillation, blood glucose monitoring in diabetic patient, endoscopy and ultrasound for gastroenterology, seizure detection devices, diagnosis of cancer with computational histopathology, and imaging-based diagnosis. In dentistry, AI has been reported in recent years for the early diagnosis of dental caries in bitewing radiograph; for classification of the early childhood caries status in children; for detection of dental plaque on deciduous teeth; for the processing or application of zirconia crowns/restorations in dentistry; for endodontic procedures such as locating apical foramen, prediction of periapical pathologies, retreatment predictions, analysis of root morphologies, and detection of vertical root fractures; for the diagnosis in orthodontics specialty and its treatment planning; and in clinical decision-making process.
Machine learning (ML), as a subfield of AI, provides imperative tools for intelligent data analysis. The three main branches of ML have been developed: statistical methods, symbolic learning, and neural networks. ANN algorithms are classified into: feed-forward neural networks (e.g., single-layer perceptron, MLP, and radial basis function networks) or recurrent neural networks (e.g., competitive network and Hopfield network).
ANN have been found successful as a digital tool in solving highly complex problems of physical science by rapid data collection and processing., Its utilization is explicitly valuable in some conditions such as, when study data manifests complex interactions or when it does not satisfy the parametric assumptions, when the relationship between dependent and independent variables is not firm, when there is a large inexplicable variance in the information, or in circumstances of poorly understood theoretical basis.,
Till now, no study has been published investigating the strength of influence of orofacial anatomical risk factors such as molar relationship, incisal overjet, overbite, and lip competency on the number of affected permanent teeth with TDI using deep learning algorithm of AI statistical methods. Therefore, the present study was done to analyze the influence of various anatomical risk factors such as overjet, overbite, molar relationship, and lip competency in determining the number of traumatized teeth per affected individual using the multilayer perceptron (MLP) model of ANN of deep learning algorithm of AI. This is the first study where all the possible anatomical risk factors together have been analyzed with the help of MLP model of AI. This article would be helpful for future researches in exploring the various other aspects of dental traumatology using the deep learning algorithm statistics of AI.
| Materials and Methods|| |
The ethical approval for this study (D. N. 1030/FM) was provided by the Institutional Ethics Committee (IEC), Faculty of Medicine of Aligarh Muslim University, Uttar Pradesh, India on July 13, 2018. This was the cross-sectional study comprised of school-going-children aged 12 and 15 year selected randomly from ten schools from different locations of a city. The participating schools’ authorities were contacted and written informed consent was obtained after explaining the present study’s objectives and significances. Likewise, the informed consent was also obtained from the school children’s parents/caregivers. The sample size was determined by the following formula as:
(where, n = sample size; Z = z statistics for given level of confidence = 1.96 (for 95% Confidence Interval); p = expected prevalence = 39.5%; d = precision = 5.0%). After round off, the total sample size was calculated as 1000. The multistage cluster sampling technique was adopted for selecting the study population. A suitable schedule for conducting the study procedures in each school were discussed and finalized with schools’ authorities.
The age of participants was determined by seeing the children’s school identity card/school record or by observing the eruption status of dentition. The eligible participants were selected after considering the inclusion and exclusion criteria. Systemically healthy children without acute illness, willing children with consent from parents, aged between 12 and 15 year were included into the study. Children without informed consent, with acute illness, and undergoing/underwent orthodontic therapy made the exclusion criteria.
The intraexaminer reproducibility, accuracy, and consistency were assessed by kappa statistic and duplicate examination as per WHO guidelines (Oral health surveys-basic methods, WHO 2013). The study data were collected by structured interview and dental examination using the self-prepared structured pro forma by the single investigator. Pro forma consisted of two sections where first section was for recording demographic data such as name, age, gender, school name, area, while second section comprised questions pertaining to dental trauma history, oral examination findings, risk factors (molar relationship, incisal overjet, overbite, and lip competency). Dental examination and structured interview of the study participants were performed in school hours within the school premises in relaxed environment where adequate illumination was available.
Strict infection control precautions were followed during entire study. Sterilized clean diagnostic instruments were used for dental examination. This study was carried out as per Declaration of Helsinki. TDI was recorded using Ellis and Davey classification. Ellis class VI fracture was not considered in study as no radiograph facility was feasible in the schools. Overjet was measured using the Community Periodontal Index of Treatment Needs (CPITN) probe as described by WHO. Overbite was also measured using CPITN probe. Lip competence was assessed as described by Burden.
The collected data were entered in Microsoft Excel sheet (version 2010). First descriptive analysis of all the dependent and independent variables was performed using SPSS software version 20. The association and strength of impact of orofacial anatomic risk factors were analyzed by MLP model of ANN of deep learning algorithm using the IBM SPSS Modeler software version-18 (IBM Corp., Armonk, N.Y., USA).
[Table 1] describes the MLP algorithm, where the “input layer” comprised of independent variables, that is, risk factors like molar relationship, incisal overjet, overbite, and lip competency). There was one “hidden layer.” The “output layer” consisted of dependent variables, that is, number of traumatized teeth with dental injuries. Hyperbolic tangent and softmax were chosen as the activation function for the hidden and output layer, respectively, in the MLP system.
|Table 1 Description of neural network of multilayer perceptron (MLP) used during present study|
Click here to view
| Results|| |
In this study, the total number of traumatized teeth was found as 164 in 135 affected children with TDI out of 1000 school children. [Table 2] depicts that majority of study participants with TDI had trauma in single tooth (80%) followed by two teeth (18.5%) and three teeth (1.5%). [Table 3] shows the frequency distribution of orofacial anatomical risk factors in study participants with TDI. Majority of affected children with dental injury had accentuated overjet (55.6%) as compared to normal overjet (<3.5 mm). Mostly, traumatized children with dental injuries had class-I molar relation (51.1%) followed by class-II division 1 (43.7%), class-II division 2 (4.4%), while class-III malocclusion showed minimum TDIs (0.7%). Majority of participants (51.1%) with TDIs showed reduced overbite as compared to accentuated overbite. Majority of affected children with TDI were found with competent lips (57%) as compared to incompetent lips (43%) [Table 3].
|Table 2 Frequency distribution of number of traumatized teeth per subject|
Click here to view
|Table 3 Distribution of anatomical risk factors in study participants with TDI|
Click here to view
[Table 1] and [Figure 1] depicting the input layers, hidden layers, and output layers of the MLP model where the estimated parameter indicating the predictors (input layers: molar relationship, incisal overjet, overbite, and lip competency) and predicted layers (hidden and output layer: number of traumatized teeth in with TDI). [Figure 1] demonstrating the nodes of all the layers connected by synaptic weight. [Table 4] and [Figure 2] showing strength of independent variables (risk factors) in the occurrence of TDI in number of teeth in terms of normalized importance. [Figure 2] shows that overbite (100%) was the strongest risk factor for the occurrence of TDI in number of teeth of affected participants followed by molar relationship (90.2%), overjet (87.7%), and the weakest risk factor was lip competency (57.2%).
|Figure 1 Structure of the MLP neural network to predict the association and strength pattern of risk factors. The three-layered fully connected multilayer perceptron comprising of 10 input nodes, four hidden nodes, and three output nodes.|
Click here to view
|Table 4 Prediction of strength of risk factors (independent variables) by MLP model in terms of normalized Importance|
Click here to view
|Figure 2 Strength of orofacial anatomical risk factors in terms of normalized importance shown by MLP model of deep learning algorithm.|
Click here to view
| Discussion|| |
In this research, anterior overbite (<3 mm), class-I molar relation, accentuated overjet (>3 mm), and competent lips were found in greatest frequency in affected children with TDIs.
In the present study, the strength of impact of oral risk factors (molar relation, overjet, lip competency, and overbite/openbite) on the causation of TDI in the number of teeth was analyzed using the MLP system, that is, a type of ANN technique of deep learning algorithm. To date, MLP method has not been used in any study analyzing the effect of risk factors in various parameters of TDI. The results of the present study described the complex associations of risk factors with the number of traumatized teeth in a quite simple and understanding way and hence, the overbite as an oral risk factor was found more strongly associated with the number of traumatized teeth with dental injury followed by molar relationship, incisal overjet, and lip competency.
The reason for getting more dental trauma with lowered overbite (<3 mm, open bite tendency) might be due to the lack of protective effect from the overlapping of maxillary anterior teeth with mandibular anterior teeth in occlusion. As the overbite increases more (deep bite tendency, >3.5 mm) there would be more protective effect from the vertical overlapping of anterior teeth (greater surface area will be covered). Overbite as a risk factor for TDI has been investigated in very few studies where increased overbite was contributing to TDI.-
The accentuated overjet was found in majority of participants with TDI. Hence, overjet makes the anterior teeth more vulnerable to dental trauma due to their protrusion. Similar findings have been reported by previous studies. One study found no association among TDI, overbite, and overjet size.
In the present study, the higher occurrence of TDI in children with Angle’s class-I molar relation may be because it is more common as compared to its counter parts. Class-1 malocclusion is more commonly found. In class-I type 2 malocclusion (Dewey’s modification, 1935), class-I molar relationship with proclined maxillary incisors (i.e., increased overjet) is found to be the reason of finding more TDI in such individuals.
Similar findings were reported in previous studies.,
Kramer et al. reported that molar relation had strong impact on TDI occurrence even after correcting the accentuated overjet and thus, demonstrated the independent impact of such class-1 malocclusion in anteroposterior direction.
Majority of TDIs were found in the participants with adequate lip coverage, and its possible reason might be the higher magnitude of impact of blow during traumatic incident dominates over the protective cushioning effect of soft tissue of lip coverage. Lip tissue get lacerated due to traumatic injury over orofacial region along with dental trauma which may be the possible reason of the present finding. Similarly, Traebert et al. found no association between TDI and inadequate lip coverage.
The varying results of the studies might be due to the interplay among oral predisposing factors and behavioral factors and environmental factors for TDI. Nguyen et al. reported that age and gender are the confounding factors which influence the association between dental trauma and overjet. Thus, it seems that studies which do not adjust the confounders may show biased associations.
Mostly, studies have investigated the association among TDI, overjet, and lip coverage using regression analysis.- One study reported relation among TDIs, overjet, molar relation, and lip coverage using the Mantel–Haenszel common odds ratio method. Very few studies evaluated the overbite as risk factor among other factors.
In dental literature, age and gender have been the confounding factors for TDIs and have been mostly analyzed by conventional regression statistical models. Confounders are considered as extraneous variables that distort the apparent relationship between input (independent) and output (dependent) variables and hence lead to erroneous conclusions.
AI, especially ML, have evolved rapidly in respect of data analysis intelligently. These ML methods make minimal assumptions regarding the data generating systems and they can be helpful even when the data are collected without a carefully controlled experimental design and also when the complicated nonlinear interactions are present. It aids in the prediction of the patterns/relations in cumbersome data by employing general purpose learning algorithms. Generally, ANN model consists of three layers of neurons: input (receives information), hidden (responsible for extracting patterns, perform most of internal processing), and output (produces and presents final network outputs). Its processing units (nodes or neurons) are interconnected through synaptic weights that permit signals to travel via network. Feed-forward networks are the common type of neural networks in medical applications. These can be single layered (e.g., Adaptive Linear Neuron) or multilayered (e.g. MLP).
The strength of the current study was the application of AI in the evaluation of impact of different anatomical risk factors on the number of teeth affected with TDIs in children and adolescents. Moreover, it provided the different perspective and valuable information with interpretation in simple ways about the interaction of various orofacial risk factors related to TDIs by employing the advanced neural network of deep learning algorithm of AI. As such, there are no limitations of the present study, yet relatively larger sample size and other age groups (e.g., young adults and older adults) could have been included. Hence, future researches related to dental traumatology using AI algorithm can include the larger sample size with other age groups.
| Conclusion|| |
In the present study, by using the MLP model of deep learning algorithm of AI statistical method, overbite was found as the strongest anatomical risk factor in determining the number of traumatized teeth per affected individual as compared to molar relationship, overjet, and lip competence.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Lam R. Epidemiology and outcomes of traumatic dental injuries: a review of the literature. Aust Dent J 2016;61(Suppl1):4–20.
Andreasen JO, Andreasen FM. Textbook and Color Atlas of Traumatic Injuries to the Teeth. 3rd ed. Copenhagen, Denmark: Munksgaard; 1994.
Petti S, Glendor U, Andersson L. World traumatic dental injury prevalence and incidence, a meta-analysis-One billion living people have had traumatic dental injuries. Dent Traumatol 2018;34:71–86.
Andersson L. Epidemiology of traumatic dental injuries. J Endod 2013;39(3 Suppl):S2–S5.
Levin L, Zadik Y. Education on and prevention of dental trauma: it’s time to act! Dent Traumatol 2012;28:49–54.
Nicolau B, Marcenes W, Sheiham A. The relationship between traumatic dental injuries and adolescent’s development along the life course. Community Dent Oral Epidemiol 2003;31:306–313.
Baxevanos K, Topitsoglou V, Menexes G, Kalfas S. Psychosocial factors and traumatic dental injuries among adolescents. Community Dent Oral Epidemiol 2017;45:449–457.
Miles J, Shelvin M. Applying Regression and Correlation. London: Sage Publication; 2001.
Tu YK, Kellett M, Clerehugh V, Gilthorpe MS. Problems of correlations between explanatory variables in multiple regression analyses in the dental literature. Br Dent J 2005;199:457–461.
Tu JV. Advantages and disadvantages of using artificial neural networks versus logistic regression for predicting medical outcomes. J Clin Epidemiol 1996;49:1225–1231.
Hashimoto DA, Witkowski E, Gao L, Meireles O, Rosman G. Artificial intelligence in anesthesiology: current techniques, clinical applications, and limitations. Anesthesiology 2020;132:379–394.
Briganti G, Le Moine O. Artificial intelligence in medicine: today and tomorrow. Front Med (Lausanne) 2020;7:27.
Lee S, Oh SI, Jo J, Kang S, Shin Y, Park JW. Deep learning for early dental caries detection in bitewing radiographs. Sci Rep 2021;11:16807.
Karhade DS, Roach J, Shrestha P et al.
An automated machine learning classifier for early childhood caries. Pediatr Dent 2021;43:191–7.
You W, Hao A, Li S, Wang Y, Xia B. Deep learning-based dental plaque detection on primary teeth: a comparison with clinical assessments. BMC Oral Health 2020;13;20:141.
Luo F, Hong G, Wan Q. Artificial intelligence in biomedical applications of zirconia. Front Dent Med 2021;2:689288.
Boreak N. Effectiveness of artificial intelligence applications designed for endodontic diagnosis, decision-making, and prediction of prognosis: a systematic review. J Contemp Dent Pract 2020;21:926–934.
Khanagar SB, Al-Ehaideb A, Vishwanathaiah S et al.
Scope and performance of artificial intelligence technology in orthodontic diagnosis, treatment planning, and clinical decision-making – a systematic review. J Dent Sci 2021;16:482–492.
Kononenko I. Machine learning for medical diagnosis: history, state of the art and perspective. Artif Intell Med 2001;23:89–109.
Kayri M. Data optimization with multilayer perceptron neural network and using new pattern in decision tree comparatively. J Comput Sci 2010;6:606–612.
Shahid N, Rappon T, Berta W. Applications of artificial neural networks in health care organizational decision-making: a scoping review. PLoS One 2019;14: e0212356.
Scarborough D, Somers MJ. Neural Networks in Organizational Research: Applying Pattern Recognition to the Analysis of Organizational Behavior. Washington, DC: American Psychological Association; 2006.
Petersen , Erik Poul, Baez , Ramon J. World Health Organization. Oral health surveys: basic methods, 5th ed. World Health Organization. 2013.
WHO. Oral Health Surveys: Basic methods, 4th ed. Geneva: World Health Organization. 1997.
Burden DJ. An investigation of the association between overjet size, lip coverage, and traumatic injury to maxillary incisors. Eur J Orthod 1995;17:513–517.
Ahlawat B, Kaur A, Thakur G, Mohindroo A. Anterior tooth trauma: a most neglected oral health aspect in adolescents. Indian J Oral Sci 2013;4:31–7. [Full text]
El-Kalla IH, Shalan HM, Bakr RA. Impact of dental trauma on quality of life among 11–14 years schoolchildren. Contemp Clin Dent 2017;8:538–44.
] [Full text]
Järvinen S. Traumatic injuries to upper permanent incisors related to age and incisal overjet. A retrospective study. Acta Odontol Scand 1979;37:335–338.
Borzabadi-Farahani A, Borzabadi-Farahani A. The association between orthodontic treatment need and maxillary incisor trauma, a retrospective clinical study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2011;112:e75–80.
Sulieman AG, Awooda EM. Prevalence of anterior dental trauma and its associated factors among preschool children aged 3-5 years in Khartoum City, Sudan. Int J Dent 2018;2018:1.
Faus-Damiá M, Alegre-Domingo T, Faus-Matoses I, Faus-Matoses V, Faus-Llácer VJ. Traumatic dental injuries among schoolchildren in Valencia, Spain. Med Oral Patol Oral Cir Bucal 2011;16:e292–5.
Kaur N, Hiremath SS. Prevalence of traumatic injuries to permanent anterior teeth among 8-15 years old government and private school children in Bangalore city. J Indian Assoc Public Health Dent 2011;9:357–63.
Kramer PF, Pereira LM, Ilha MC, Borges TS, Freitas MPM, Feldens CA. Exploring the impact of malocclusion and dentofacial anomalies on the occurrence of traumatic dental injuries in adolescents. Angle Orthod 2017;87:816–823.
Traebert J, Peres MA, Blank V, Bo¨ell Rda S, Pietruza JA. Prevalence of traumatic dental injury and associated factors among 12-year-old school children in Florianopolis, Brazil. Dent Traumatol 2003;19:15–18.
Glendor U. Aetiology and risk factors related to traumatic dental injuries: a review of the literature. Dent Traumatol 2009;25:19–31.
Nguyen QV, Bezemer PD, Habets L, Prahl-Andersen B. A systematic review of the relationship between overjet size and traumatic dental injuries. Eur J Orthod 1999;21:503–515.
Hoyte T, Kowlessar A, Ali A, Bearn D. Prevalence and occlusal risk factors for fractured incisors among 11–12-year-old children in the Trinidad and Tobago Population. Dent J (Basel) 2020;8:25.
Gupta M, Kumar S, Kaur J, Gupta S, Bansal V, Dwiedi A., Prevalence, risk factors, and treatment needs of traumatic dental injuries of anterior teeth among 11-15 year old children attending government and private schools of Bhopal city, India. J Adv Oral Res 2016;7:32–9.
Gupta S, Kumar-Jindal S, Bansal M, Singla A. Prevalence of traumatic dental injuries and role of incisal overjet and inadequate lip coverage as risk factors among 4-15 years old government school children in Baddi-Barotiwala Area, Himachal Pradesh, India. Med Oral Patol Oral Cir Bucal 2011;16:e960–e965.
Patel MC, Sujan SG. The prevalence of traumatic dental injuries to permanent anterior teeth and its relation with predisposing risk factors among 8-13 years school children of Vadodara city: an epidemiological study. J Indian Soc Pedod Prev Dent 2012;30:151–7.
Shulman JD, Peterson J. The association between incisor trauma and occlusal characteristics in individuals 8-50 years of age. Dent Traumatol 2004;20:67–74.
Sarker IH. AI-driven cybersecurity: an overview, security intelligence modeling and research directions. SN Comput Sci 2021;2:173.
Bzdok D, Altman N, Krzywinski M. Statistics versus machine learning. Nat Methods 2018;15:233–234.
Silva IN, Hernane Spatti S, Flauzino RA. Artificial neural network architectures and training processes. In: Artificial Neural Networks: A Practical Course. Cham, Switzerland: Springer International Publishing; 2017.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]