CBD2023: A Hypercomplex Bangla Handwriting Character Recognition Data for Hierarchical Class Expansion

Object recognition technology has made significant strides, but recognizing handwritten Bangla characters (including symbols, compound forms, etc.) remains a challenging problem due to the prevalence of cursive writing and many ambiguous characters. The complexity and variability of the Bangla script and individual's unique handwriting styles make it difficult to achieve satisfactory performance for practical applications, and the best existing recognizers are far less effective than those developed for English alpha-numeric characters. Compared to other major languages, there are limited options for recognizing handwritten Bangla characters. This research has described a new dataset to improve the accuracy and effectiveness of handwriting recognition systems for the Bengali language spoken by over 200 million people worldwide. This dataset aims to investigate and recognize Bangla handwritten characters, focusing on enlarging the recognized character classes. To achieve this, a new challenging dataset for handwriting recognition is introduced, collected from numerous students' handwriting from two institutions.

Psychometric assessment of the Bangla version of the Bergen Social Media Addiction Scale

Recently, researchers have raised their concern about problematic engagement in social media use that significantly impacted users’ mental health and daily lives. Therefore, it is important to have a psychometrically sound assessment tool to assess social media addiction. The present study aimed to assess the reliability and validity of the Bangla version of the Bergen Social Media Addiction Scale (BSMAS) for assessing social media addiction among young Bangladeshi adults. In this study, we utilized secondary data that contained information from 577 Bangladeshi university students. Exploratory factor analysis explored a single latent factor, and confirmatory factor analyses supported this structure. Discrimination indices of items in both classical test theory (CTT) and item response theory (IRT) approach suggested that items could discriminate between low scorers and high scorers in this scale. This scale has good internal consistency, composite, and IRT reliability. Multigroup CFA and differential item functioning bias results suggested this scale would be assessed the same construct across gender and usage duration groups (5 h and more vs less than 5 h). Network analysis results suggested relapse following salience as the core symptoms of social media addiction among young Bangladeshi adults. Overall, results suggested the Bangla BSMAS as a psychometrically sound tool to assess symptoms of social media addiction among young Bangladeshi adults. This scale has practical utility to mental health practitioners as this scale provides information about the core symptoms of social media addiction.

The psychometric properties of the Bangla Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5): preliminary reports from a large-scale validation study

BACKGROUND

The Posttraumatic Stress Disorder Checklist (PCL-5) is the most widely used screening tool in assessing posttraumatic stress disorder symptoms, based on the Diagnostic and Statistical Manual of Mental disorders (DSM-5) criteria. This study aimed to evaluate the psychometric properties of the newly translated Bangla PCL-5.

METHODS

A cross-sectional survey was carried out among 10,605 individuals (61.0% male; mean age: 23.6 ± 5.5 [13-71 years]) during May and June 2020, several months after the onset of the COVID-19 outbreak in Bangladesh. The survey included the Bangla PCL-5 and the PHQ-9 depression scale. We used confirmatory factor analysis to test the four-factor DSM-5 model, the six-factor Anhedonia model, and the seven-factor hybrid model.

RESULTS

The Bangla PCL-5 displayed adequate internal consistency (Cronbach's alpha = 0.90). The Bangla PCL-5 score was significantly correlated with scores of the PHQ-9 depression scale, confirming strong convergent validity. Confirmatory factor analyses indicated the models had a good fit to the data, including the four-factor DSM-5 model, the six-factor Anhedonia model, and the seven-factor hybrid model. Overall, the seven-factor hybrid model exhibited the best fit to the data.

CONCLUSIONS

The Bangla PCL-5 appears to be a valid and reliable psychometric screening tool that may be employed in the prospective evaluation of posttraumatic stress disorder in Bangladesh.

Reliability and validity assessment of the Mental Health Continuum - Short Form for Bangladeshi adults

There is a shortage of psychological tools to assess the mental well-being of Bangladeshi people covering the hedonic and the eudaimonic aspects of well-being. This study aimed to assess the reliability and validity of the Mental Health Continuum Short Form as this instrument assesses both the hedonic and the eudaimonic aspects of well-being. This study utilized online survey data (n = 361). Item analysis results showed that this scale had good discrimination indices, internal consistency reliability, and acceptable average inter-item correlations. Confirmatory factor analysis lent support to the three-factor correlated model with good model fits and sufficient factor loadings. Results from the multidimensional graded response model also supported the efficiency of this scale to assess the mental well-being of Bangladeshi adults. This scale had moderate correlations with perceived stress and anxiety. Overall, findings suggested this scale as a reliable and valid tool for assessing mental well-being. This scale would help mental health practitioners to assess mental well-being and formulate relevant interventions.

Validation of the Bangla WHO-5 Well-being Index

BACKGROUND

Subjective wellbeing in terms of objective outcome can be useful to determine the level of progress in clinical practice as well as research studies in Bangladesh. Besides, cultural understanding of well-being for Bangladeshi population is also equally important to report. A valid Bangla version of the five-item WHO Well-being Index can be a suitable measure to achieve the purposes. Therefore, the present study aimed at validating the WHO-5 Well-being Index for general population in Bangladesh.

METHODS

After following the standard procedures for translation, back-translation, and committee translation, the initial Bangla version of the scale was developed and pretested. Based on the feedback during pretesting, a slight modification was made and the final version was developed. This final version was administered to 269 participants of different socioeconomic backgrounds to find out the reliability and validity of the scale from March 2019 to May 2019. The data analysis was conducted using SPSS 24.

RESULTS

The scale demonstrated acceptable internal consistency (α = 0.754) and test-retest reliability (r = 0.713), divergent validity (r = -0.443, p < 0.01 with the Bangla version of Perceived Stress Scale-10) and convergent validity (r = 0.542, p < 0.01 with the Bangla version of Warwick-Edinburgh Mental Well-Being Scale). The data also yielded one-factor structure for the scale in exploratory factor analysis explaining 38.68% of total variance. The factor-structure was further supported in the confirmatory factor analysis (χ² = 295.852, χ²/df = 2.017, RMSEA = 0.062, CFI = 0.986, TLI = 0.964, and SRMR = 0.0255).

CONCLUSION

The findings suggested the Bangla version of the WHO-5 Well-being Index is a psychometrically valid and reliable tool for general adult population in Bangladeshi when it comes to measuring subjective well-being both in clinical practice and research studies.

Validation of the Bangla Mindful Attention Awareness Scale

This study tested the reliability, validity and factor structure of the Bangla translated Mindful Attention Awareness Scale (MAAS). Three scales - the Bangla MAAS, the short form of health 36 (SF-36) and the Internet Addiction Test (IAT) - were applied to 519 university students (51.4% female). Ninety-two participants were retested over a two weeks period to examine test-retest reliability. Consistent with previous studies, the results of exploratory and confirmatory factor analyses showed a single factor solution for the Bangla MAAS (χ2/df=222.243/90=2.47, CFI=0.93, NFI=0.87, and RMSEA=0.053). The temporal stability and internal consistency was also satisfactory (Cronbach's alpha 0.85). The Bangla MAAS was significantly and positively associated with SF-36 Mental and Physical health components and negatively with IAT scores. Additionally, MAAS scored significantly varied with the presence of physical illness and different living conditions, suggesting the validity of the tool. The Bangla version of the MAAS is, therefore, a valid and reliable tool to measure mindfulness among young Bangladeshi adults.

Locality and Word Order in Active Dependency Formation in Bangla

Research on filler-gap dependencies has revealed that there are constraints on possible gap sites, and that real-time sentence processing is sensitive to these constraints. This work has shown that comprehenders have preferences for potential gap sites, and immediately detect when these preferences are not met. However, neither the mechanisms that select preferred gap sites nor the mechanisms used to detect whether these preferences are met are well-understood. In this paper, we report on three experiments in Bangla, a language in which gaps may occur in either a pre-verbal embedded clause or a post-verbal embedded clause. This word order variation allows us to manipulate whether the first gap linearly available is contained in the same clause as the filler, which allows us to dissociate structural locality from linear locality. In Experiment 1, an untimed ambiguity resolution task, we found a global bias to resolve a filler-gap dependency with the first gap linearly available, regardless of structural hierarchy. In Experiments 2 and 3, which use the filled-gap paradigm, we found sensitivity to disruption only when the blocked gap site is both structurally and linearly local, i.e., the filler and the gap site are contained in the same clause. This suggests that comprehenders may not show sensitivity to the disruption of all preferred gap resolutions.

Psychometric properties of two measures of childhood internalizing problems in a Bangladeshi sample

OBJECTIVES

In order to assist mental health services in developing countries, a key issue is the availability of psychometrically sound, brief, and cost-effective measures that have been tested within the relevant context. The present study was designed to evaluate within a young Bangladeshi population, the psychometric properties of two widely used Western measures of internalizing distress in young people: the short form of the Spence Children's Anxiety Scale and the Short Moods and Feelings Questionnaire.

METHOD

The sample included 1,360 children and adolescents aged 9-17 years (M = 12.3 years, SD = 2.12) recruited from six districts of Bangladesh, including both community and emotionally at-risk participants. A total of 179 children were re-tested on the measures within 3-4 weeks.

RESULTS

Confirmatory factor analyses showed single-factor structures for both scales in the total sample and in both community and at-risk participants separately. Multiple group analyses across gender and age-group within the at-risk and community samples showed that the single-factor structure was suitable regardless of subgroup. Analyses also indicated acceptable internal consistency, test-retest reliability and construct validity for both scales.

CONCLUSION

The two measures show promise as brief, reliable, and valid instruments for the assessment of internalizing distress among young people from Bangla-speaking communities.

PRACTITIONER POINTS

Positive clinical implications: These two measures of internalizing distress in young people showed solid psychometric properties within samples collected from various parts of Bangladesh. The measures can therefore be used to assess anxiety and depression in Bangla-speaking youth. These measures should be of value in both clinical settings and at a community level to assess the need for services. Cautions and limitations: Resource limitations did not allow comparison against diagnostic criteria and therefore cut-off scores to indicate clinical status among Bangladeshi youth will require further research.

Psychometric properties of the Children's Revised Impact of Events Scale (CRIES) with Bangladeshi children and adolescents

Identification of possible cases suffering post-traumatic stress disorder (PTSD) is important, especially in developing countries where traumatic events are typically prevalent. The Children's Revised Impact of Events Scale is a reliable and valid measure that has two brief versions (13 items and 8 items) to assess reactions to traumatic events among young people. The current study evaluated the psychometric properties of both versions of the CRIES in a sample of 1,342 children and adolescents aged 9-17 years (M = 12.3 years, SD = 2.12) recruited from six districts of Bangladesh. A sub-group of 120 children from four schools was re-tested on the measures within 3.5 weeks. Confirmatory factor analysis supported factor structures similar to those found in other studies for both versions of the CRIES. Multiple group confirmatory factor analysis showed gender and age-group differences within the sample, supporting established age and gender differences in prevalence of PTSD symptoms. Analyses also indicated moderate to excellent internal consistency and test-retest reliability and clear discriminant and convergent validity. These data support use of both the CRIES-13 and CRIES-8 to provide quick and psychometrically sound assessment of symptoms of PTSD among children and adolescents from Bangla-speaking communities.

The Internet Addiction Test: assessing its psychometric properties in Bangladeshi culture

There is growing importance of the Internet Addiction Test (IAT) in Internet addiction research around the world. Since the development of the IAT (Young, 1996, 1998), a number of validation studies have been done in various cultures. The aim of this study was to translate the instrument into Bangla and validate in Bangladeshi culture, a culture vulnerable to Internet addiction. A total of 177 Internet users (77 females and 100 males) participated in the study. Exploratory factor analysis (EFA) of the data from 172 participants (who provided complete responses) identified a four factor structure of the IAT with 18 items. The four factors namely 'Neglect of duty', 'Online dependence', 'Virtual fantasies', and 'Privacy and self-defense' together explained 55.68% of the total variance. Problematic (moderate/excessive) users on the IAT scored, on average, higher on each of the four IAT factors as compared to average or non-problematic (minimal) users consistently across genders. The IAT and its factors showed good internal consistency (Cronbach's α=.89 for the IAT, and .60-.84 for the factors), strong convergent and discriminant validity. Thus, the Bangla version IAT appears to be valid and reliable and therefore may be used in further research on Internet addiction in the country.