Exhaustive Chi-squared Automatic Interaction Detector Research Articles

Comparison of machine learning algorithms and multiple linear regression for live weight estimation of Akkaraman lambs

This study was designed to predict the post-weaning weights of Akkaraman lambs reared on different farms using multiple linear regression and machine learning algorithms. The effect of factors the age of the dam, gender, type of lambing, enterprise, type of flock, birth weight, and weaning weight was analyzed. The data was collected from a total of 25,316 Akkaraman lambs raised at multiple farms in the Çiftlik District of Niğde province. Comparative analysis was conducted by using multiple linear regression, Random Forest, Support Vector Machines (and Support Vector Regression), Extreme Gradient Boosting (XGBoost) (and Gradient Boosting), Bayesian Regularized Neural Network, Radial Basis Function Neural Network, Classification and Regression Trees, Exhaustive Chi-squared Automatic Interaction Detection (and Chi-squared Automatic Interaction Detection), and Multivariate Adaptive Regression Splines algorithms. In this study, the test dataset was divided into five layers using the K-fold cross-validation method. The performance of models was compared using performance criteria such as Adjusted R-squared (Adj-\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${R}^{2}$$\\end{document}), Root Mean Square Error (RMSE), Mean Absolute Deviation (MAD), and Mean Absolute Percentage Error (MAPE) by utilizing test populations in the predicted models. Additionally, the presence of low standard deviations for these criteria indicates the absence of an overfitting problem. \\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${R}^{2}$$\\end{document}The comparison results showed the Random Forest algorithm had the best predictive performance compared to other algorithms with Adj-\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${R}^{2}$$\\end{document}, RMSE, MAD, and MAPE values of 0.75, 3.683, 2.876, and 10.112, respectively. In conclusion, the results obtained through Multiple Linear Regression for the live weights of Akkaraman lambs were less accurate than the results obtained through artificial neural network analysis.

Use of data mining algorithms to predict body weight of Dorper sheep Breed in two selected villages of Lepelle-Nkumpi local municipality in Limpopo province, South Africa

Sheep farmers in village areas find it difficult to estimate body weight without the weighing tools. The best fitted model for estimation of sheep body weight is not conclusive. Classification and regression tree (CART), Chi-square automatic interaction detection (CHAID), Exhaustive chi-square automatic interaction detection (Ex-CHAID) and Multivariate adaptive regression spline (MARS) were used to predict body weight (BW) of 306 Dorper sheep (female = 244 and male = 62) aged 1 to 2 years from linear body measurement traits such as rump height (RH), rump length (RL), heart girth (HG), withers height (WH), and body length (BL). Goodness of fit criteria's such as Pearson's correlation coefficients (r), coefficient of determination (R2), Akaike's information criterion AIC and adjusted R2 (Adj.R2) were computed for data mining algorithms performance assessment. The results indicated that MARS was the best data mining algorithm with r of 0.979, R2 of 0.958, Adj.R2 of 0.957, RMSE of 2.503 and AIC of 422.25. The MARS model provided the most appropriate predictive capability in the prediction of body weight for Dorper sheep with HG, WH, RH, BL, and sex (male) as the important traits. As a result, MARS data mining algorithm might be used to identify factors that might be vital to improve body weight of Dorper sheep for breeding programs.

Forest harvest operations productivity forecasting: a decisions tree approach

ABSTRACT In recent years, machinery used in forest harvesting operations has incorporated the ability to collect data during the harvesting operation automatically. The processing of these data allows for obtaining new perspectives on the harvest characteristics. In this sense, it is that the development of predictive models for harvest productivity is approached by processing the automatically retrieved data. In turn, these new models pave the way to develop new tools for operations management decision-making processes, providing a data-driven approach. In this case, forest productivity is analyzed based on different harvesting operational configurations defined by stand, trees, species, operators, shifts, etc. which make it possible to adequately predict what the wood-harvested volume will be, and thus, synchronize the rest of the supply chain logistics operations. The data processing is done through decision tree methods. Different methods of decision trees based on exhaustive Chi-square automatic interaction detection (CHAID), recursive binary partition, and conditional inference based that also uses binary recursive partition are tested. The results show that decision recursive binary partition methods tend to model the entire spectrum of the target variable in a more balanced manner. While exhaustive CHAID-based methods tend to be more accurate in global terms but more unbalanced. As a general comment of the method, global confusion matrices are around 50% accurate, and some operational configurations and productivity classes are predicted with almost 90% accuracy.

Predictors of Adolescent Intimate Partner Sexual Violence Victimization: Patterns of Intersectional Social Positions in a Statewide, School-based Sample.

Intimate partner sexual violence (IPSV) is a substantial public health issue faced by youth, with disparities along lines of race, gender, and sexual orientation. Using an intersectional framework, the current study describes the prevalence of self-reported IPSV victimization across intersecting social positions among adolescents. High school students who completed the 2019 Minnesota Student Survey were asked whether they had ever experienced IPSV. Using exhaustive Chi-Square Automatic Interaction Detection (eCHAID), responses were modeled using five indicators: racial/ethnic identity, transgender/genderqueer/genderfluid identity, sex assigned at birth, sexual orientation, and school location. The prevalence of IPSV victimization within each end group identified by the eCHAID varied from 2.6% to 32.0%. The highest prevalence of IPSV among all students was observed in subgroups defined by multiple marginalized social positions; for example, 32.0% among lesbian, gay, bisexual, queer, pansexual, or questioning (LGBQ+) students of color who were transgender, genderqueer, genderfluid or unsure of their gender, and assigned male at birth; 22.8% among LGBQ+ students assigned female at birth and enrolled in a Greater Minnesota (outside of the Twin Cities) school; and 22.2% among Native+ or missing race LGBQ+ students assigned female at birth and enrolled in a metropolitan school. Within the eCHAID decision tree, several subsamples of students were split into two more subsamples with a twofold or greater statistically significant difference in IPSV victimization prevalence between them. Findings from this study highlight notable disparities in the prevalence of IPSV victimization among Minnesota youth and demonstrate the importance of detailing the distribution of interpersonal violence outcomes across youth with multiple marginalized identities. Findings suggest that disparities reported along one categorical axis only may miss important nuances in how risk for IPSV is distributed. Population health researchers should utilize methods that allow for the explication of complex intersections that characterize individuals' social positions in estimating the prevalence of sexual violence.

Bright diffusion sign: A sensitive and specific radiologic biomarker for multinodular and vacuolating neuronal tumor

Background and purposeAccurate differentiation between multinodular and vacuolating neuronal tumor (MVNT) and dysembryoplastic neuroepithelial tumor (DNET) is important for treatment decision-making. We aimed to develop an accurate radiologic diagnostic model for differentiating MVNT from DNET using T2WI and diffusion-weighted imaging (DWI). Materials and methodsA total of 56 patients (mean age, 47.48±17.78 years; 31 women) diagnosed with MVNT (n = 37) or DNET (n = 19) who underwent brain MRI, including T2WI and DWI, were included. Two board-certified neuroradiologists performed qualitative (bubble appearance, cortical involvement, bright diffusion sign, and bright apparent diffusion coefficient [ADC] sign) and quantitative (nDWI and nADC) assessments. A diagnostic tree model was developed with significant and reliable imaging findings using an exhaustive chi-squared Automatic Interaction Detector (CHAID) algorithm. ResultsIn visual assessment, the imaging features that showed high diagnostic accuracy and interobserver reliability were the bright diffusion sign and absence of cortical involvement (bright diffusion sign: accuracy, 94.64 %; sensitivity, 91.89 %; specificity, 100.00 %; interobserver agreement, 1.00; absence of cortical involvement: accuracy, 92.86 %; sensitivity, 89.19 %; specificity, 100.00 %; interobserver agreement, 1.00). In quantitative analysis, nDWI was significantly higher in MVNT than in DENT (1.52 ± 0.34 vs. 0.91 ± 0.27, p < 0.001), but the interobserver agreement was fair (intraclass correlation coefficient = 0.321). The overall diagnostic accuracy of the tree model with visual assessment parameters was 98.21 % (55/56). ConclusionThe bright diffusion sign and absence of cortical involvement are accurate and reliable imaging findings for differentiating MVNT from DNET. By using simple, intuitive, and reliable imaging findings, such as the bright diffusion sign, MVNT can be accurately differentiated from DNET.

Yield prediction models of organic oil rose farming with agricultural unmanned aerial vehicles (UAVs) images and machine learnaing algorithms

The alternative solution for predicting yields involves leveraging the latest technological advancements in unmanned aerial vehicles (UAVs) imagery. This approach aims to empower agricultural systems operating in nature-dependent conditions, enabling them to optimize the utilization of limited resources effectively. This study aimed to provide the agricultural data management needed for the adoption of sustainable agriculture systems and the transition of precision to smart agriculture systems by combining technology. The study was carried out in the 2019–2020 vegetation period to determine the effect of the parameters on the yield in the areas where organic farming is grown in Isparta oil rose (Rosa Damascena Mill.) and to develop yield prediction models by combining them with spectral measurements. In the study, the phenological, morphological, and biochemical changes of the Isparta oil rose were examined periodically, and the effects of biotic and abiotic stress factors during the vegetation period were determined in the study and testing area. For the early yield prediction of organic Isparta oil rose was chosen thirteen parameters as soil moisture, pH, electrical conductivity (EC), soil temperature, and nine vegetation indices were derived from UAV-acquired multispectral images. The parameters were functioned in machine learning algorithms, which were multiple linear regression (MLR), multivariate adaptive regression splines (MARS), decision trees (chi-squared automatic interaction detector (CHAID), exhaustive chi-squared automatic interaction detector (ExCHAID), classification and regression tree (CART)), random forest (RF)) and artificial neural network (ANN), to develop the prediction models. Early period (day 69) for organic rose yield prediction models were determined to be useable as MARS 1 (0.907 R2), ExCHAID 1 (0.888 R2), CART 1 (0.931 R2), and RF 1 (0.909 R2).As a result of the study, the models can be used integrated with UAV and soil data for the early prediction of Isparta oil rose yield under organic farming systems, especially during the Agriculture 4.0 transition phase in our country. It is such that models be developed and used in different ecologies for sustainable agricultural systems.

Unaccompanied unstable housing among racially, ethnically, sexually, and gender diverse youth: Intersecting identities bearing the greatest burden.

Disparities in youth homelessness by racial/ethnic, sexual, and gender identities are well documented, though this literature lacks specificity regarding intersectional social identities of youth who are most likely to experience homelessness. Population-based cross-sectional data on youth from the 2019 Minnesota Student Survey (N = 80,456) were used to examine the relationship between parent caring and intersections of minoritized identities that experience the highest prevalence of two distinct types of unaccompanied unstable housing with expanded categories of sexual and gender identities. Exhaustive chi-square automatic interaction detection models revealed that low parent caring was the most common predictor of unaccompanied homelessness and running away, but there was important variation among youth of color at the intersection of sexual and gender identities. The findings reveal a more complex story of disparities in unaccompanied unstable housing among youth with multiple marginalized social identities and highlight the need to create culturally informed prevention and intervention strategies for parents of LGBTQ+ (lesbian, gay, bisexual, transgender, queer, and questioning) youth of color. The implications for prevention and intervention among subgroups with the highest prevalence are discussed in the context of interlocking systems of power and oppression. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Potential of preoperative fluorodeoxyglucose-positron emission tomography/computed tomography to diagnose contralateral lymph node metastases in patients with oral cancer.

To establish a decision tree using preoperative PET/computed tomography (CT) parameters for detecting contralateral lymph node metastasis (CLNM) in oral cancer patients. In total, 140 patients with a confirmed histopathological diagnosis of oral carcinoma showed fluorodeoxyglucose accumulation of contralateral lymph nodes in PET images. Of 260 lymph nodes, eight were metastatic. We compared metastatic and non-metastatic nodes using Fisher-Freeman-Halton exact and Fisher's exact and Mann-Whitney U tests. The diagnostic performance was analyzed using receiver operating characteristic curves. We established a decision tree using exhaustive chi-squared automatic interaction detection algorithm. Five PET/CT parameters were significantly different between metastatic and non-metastatic nodes: the longest ( P = 0.015) and shortest ( P = 0.023) diameter, and the maximum standardized uptake values (SUVmax) of the contralateral node ( P = 0.030), primary tumor ( P < 0.001), and ipsilateral node ( P < 0.001). The area under the curves of SUVmax of the primary tumor (0.887), ipsilateral node (0.886), and longest diameter (0.752) were the largest and these three parameters were used as predictive criteria in the decision tree. The accuracy of the decision tree was 97.7% with 100% sensitivity and 97.6% specificity. Preoperative PET/CT imaging with a decision tree has the potential to diagnose CLNM in patients with oral cancer.

Bullying Involvement at the Intersection of Gender Identity/Modality, Sexual Identity, Race, Ethnicity, and Disability: Prevalence Disparities and the Role of School-Related Developmental Assets.

Purpose: This study examines adolescents' self-reported school-based developmental assets and four intersecting social positions as they relate to prevalence of bullying involvement. Methods: Participants were 80,456 ninth and 11th grade students who participated in the 2019 Minnesota Student Survey (30.2% youth of color; 11% lesbian/gay/bisexual/pansexual/queer/questioning; 2.9% transgender/gender diverse [TGD] or gender questioning). Exhaustive Chi-square Automatic Interaction Detection analysis was used to identify school-based developmental assets (i.e., school safety, school adult support) and intersecting social positions (i.e., sexual identity; gender identity/modality; racial/ethnic identity; physical disabilities/chronic illness; and/or mental health/behavioral/emotional problems) associated with the highest prevalence of involvement as physical and relational bullies, victims, and bully-victims. Results: Adolescents with 2+ marginalized social positions who often lacked school-based developmental assets were part of nearly all the highest prevalence bullying involvement groups. TGD and gender questioning adolescents, Native American youth, and youth living with both physical disabilities/chronic illness and mental health/emotional/behavioral problems-most of whom had additional marginalized social positions and lacked school-based assets-were particularly overrepresented in high prevalence groups. For example, 31.1% of TGD or gender questioning youth of color living with both types of disabilities/health problems who did not feel strongly that school was safe reported involvement as physical bully-victims-nearly six times the sample average rate. Conclusion: Adolescents with multiple marginalized social positions and those lacking certain school-based assets-often overlapping categories-were involved in bullying at higher-than-average rates. Findings underscore the need for schools to address intersecting experiences of stigma and structural oppression that may perpetuate bullying involvement disparities.

Student Employees’ Dropout Intentions: Work Excuse and University Social Capital as Source and Solution

&lt;p style="text-align: justify;"&gt;This study seeks to investigate the relationship between dropout intent, the weekly work duration of student employees, and university social capital by analysing empirical evidence from three European countries, including Estonia, Lithuania, and Poland. This exploratory study utilised Eurostudent-VII survey data and employed cross-tabulation and exhaustive Chi-square Automatic Interaction Detection (CHAID) to achieve its objectives. Findings indicate that student employees who believe they get along well with their teachers and have more connections with fellow students to discuss subject-related issues are less likely to intend to drop out of university. In addition, the results show that students’ likelihood of abandoning their higher education increases in the presence of difficulties caused by an inapt academic programme. Regarding employment duration, for the Estonian and Lithuanian markets, there is no difference between working more than 20 hours per week or less than that with the intention of dropping out of university. In Poland, however, the disparity in working hours interacts with other factors related to social capital to explain dropout intent. These findings provide novel insights into the dropout literature by refreshing thoughts on the role of teacher-student and peer relations in the dropout intentions of student employees. In addition to reviving the relevance of university social capital, which has received too little attention lately, they have also sparked a recent debate on whether or not combining work and university actually affects the intention to drop out.&lt;/p&gt;

Exploring Bias-Based Bullying and Intersecting Social Positions as Correlates of Sexual Risk Behaviors Among Adolescents.

Purpose: The current study extends the limited body of intersectional research on adolescents' sexual health by examining experiences of bias-based bullying and multiple intersecting social positions associated with engagement in sexual risk behaviors. Methods: Participants were 14,968 sexually active 9th and 11th grade students surveyed as part of the 2019 Minnesota Student Survey (15% lesbian/gay/bisexual/queer/pansexual/questioning [LGBQ] and/or transgender/gender diverse [TGD] or gender questioning). Exhaustive Chi-square Automatic Interaction Detection analysis was used to identify experiences (i.e., bias-based bullying victimization) and intersecting social positions (i.e., sexual orientation identity; gender identity/modality; race/ethnicity; physical disabilities/chronic illness; mental health/behavioral/emotional problems) associated with the highest prevalence of three sexual risk behaviors. Results: Overall, 18% of adolescents reported 3+ sex partners in the last year, 14% reported drug/alcohol use before last sex, and 36% reported not discussing protection from sexually transmitted infections with new sexual partners. Adolescents with 2+ marginalized social positions, some of whom also experienced bias-based bullying, were part of 53% of the highest prevalence risk groups. For example, 42% of Multiracial or Latina/x/o gender questioning adolescents who identified as LGBQ reported 3+ sex partners in the last year-twice the sample average. Adolescents who were Black, American Indian/Alaska Native, Latina/x/o, Multiracial, TGD, or gender questioning were in the highest prevalence nodes across all outcomes. Conclusion: Adolescents with multiple marginalized social positions and who experience bias-based bullying engage in high-risk sexual behaviors at higher-than-average rates. Findings underscore the importance of addressing intersecting experiences of stigma to reduce high-risk sex behaviors and promote health equity among adolescents.

Interpersonal protective factors for LGBTQ+ youth at multiple intersecting social identities and positions.

Interpersonal supports are protective against multiple negative health outcomes for youth such as emotional distress and substance use. However, finding interpersonal support may be difficult for youth exposed to intersecting racism, heterosexism, and cisgenderism, who may feel they are "outsiders within" their multiple communities. This study explores disparities in interpersonal supports for youth at different sociodemographic intersections. The 2019 Minnesota Student Survey includes data from 80,456 high school students, including measures of four interpersonal supports: feeling cared about by parents, other adult relatives, friends, and community adults. Exhaustive Chi-square Automatic Interaction Detection analysis was used to examine all interactions among four social identities/positions (racialized/ethnic identity, sexual identity, gender identity, sex assigned at birth) to identify groups who report different rates of caring from each source (Bonferroni adjusted p<.05). In the overall sample, 69.24% perceived the highest level of caring ("very much") from parents, 50.09% from other adult relatives, 39.94% from friends, and 15.03% from community adults. Models identified considerable differences in each source of support. For example, more than 72% of straight, cisgender youth reported their parents cared about them very much, but youth who identified as LGBQ and TGD or gender-questioning were much less likely to report high parent caring (less than 36%) across multiple racialized/ethnic identities and regardless of sex assigned at birth. Findings highlight the importance of better understanding the ways interpersonal support might differ across groups, and underscore a need for intersectionality-tailored interventions to develop protective interpersonal supports for LGBTQ+ youth, rather than one-size-fits-all approaches.

Navigating Digital Inequality: Examining Factors Affecting Rural Customers’ Internet Banking Adoption in Post-COVID Bangladesh

As the world continues to navigate the new normal brought about by the COVID-19 pandemic, one issue that has come to the forefront is digital inequality. In Bangladesh, where a significant portion of the population resides in rural areas, the adoption of internet banking has been hindered by various factors. However, understanding these factors is crucial, especially now that digital transactions have become more important. This study aims to understand the factors influencing the adoption of internet banking services among rural customers in Bangladesh. To acquire data, a questionnaire was administered to 443 rural bank customers in the district of Barisal. The Exploratory Factor Analysis (EFA) revealed three primary factors: trust compatibility, service benefit, and access to consumer education. In addition, the research sought to determine if the identified factors, particularly access to consumer education, varied according to the occupation and income level of rural consumers. Using exhaustive Chi-squared Automatic Interaction Detection (CHAID) analysis, the findings revealed that access to consumer education differs significantly by occupation level, with business and service holders being more likely than farmers to have access to consumer education. This research contributes to the literature by providing insights into the adoption of internet banking by rural customers and informing policymakers about the special needs of this demographic.

Natural history of meningiomas: a serial volumetric analysis of 240 tumors.

The management of asymptomatic intracranial meningiomas is controversial. Through the assessment of growth predictors, the authors aimed to create the basis for practicable clinical pathways for the management of these tumors. The authors volumetrically analyzed meningiomas radiologically diagnosed at their institution between 2003 and 2015. The primary endpoint was growth of tumor volume. The authors used significant variables from the multivariable regression model to construct a decision tree based on the exhaustive Chi-Square Automatic Interaction Detection (CHAID) algorithm. Of 240 meningiomas, 159 (66.3%) demonstrated growth during a mean observation period of 46.9 months. On multivariable logistic regression analysis, older age (OR 0.979 [95% CI 0.958-1.000], p = 0.048) and presence of calcification (OR 0.442 [95% CI 0.224-0.872], p = 0.019) had a negative predictive value for tumor growth, while T2-signal iso-/hyperintensity (OR 4.415 [95% CI 2.056-9.479], p < 0.001) had a positive predictive value. A decision tree model yielded three growth risk groups based on T2 signal intensity and presence of calcifications. The median tumor volume doubling time (Td) was 185.7 months in the low-risk, 100.1 months in the intermediate-risk, and 51.7 months in the high-risk group (p < 0.001). Whereas 0% of meningiomas in the low- and intermediate-risk groups had a Td of ≤ 12 months, the percentage was 8.9% in the high-risk group (p = 0.021). Most meningiomas demonstrated growth during follow-up. The absence of calcifications and iso-/hyperintensity on T2-weighted imaging offer a practical way of stratifying meningiomas as low, intermediate, or high risk. Small tumors in the low- or intermediate-risk categories can be monitored with longer follow-up intervals.

Data-Driven Based Prediction of the Energy Consumption of Residential Buildings in Oshawa

Buildings consume about 40% of the global energy. Building energy consumption is affected by multiple factors, including building physical properties, performance of the mechanical system, and occupants’ activities. The prediction of building energy consumption is very complicated in actual practice. Accurate and fast prediction of the building energy consumption is very important in building design optimization and sustainable energy development. This paper evaluates 24 energy consumption models for 83 houses in Oshawa, Canada. The energy consumption, social and demographic information of the occupants, and the physical properties of the houses were collected through smart metering, a phone survey, and an energy audit. A total of 63 variables were determined, and based on the variable importance, three groups with different numbers of variables were selected, i.e., 26, 12, and 6 for electricity consumption; and 26, 13, and 6 for gas consumption. A total of eight data-driven algorithms, namely Multiple Linear Regression (MLR), Stepwise Regression (SR), Support Vector Machine (SVM), Backpropagation Neural Network (BPNN), Radial Basis Function Neural Network (RBFN), Classification and Regression Tree (CART), Chi-Square Automatic Interaction Detector (CHAID), and Exhaustive CHAID (ECHAID), were used to develop energy prediction models. The results show that the BPNN model has the best accuracies in predicting both the annual electricity consumption and gas consumption, with mean absolute percentage errors (MAPEs) of 0.94% and 0.94% for training and validation data for electricity consumption, and 2.63% and 0.16% for gas consumption, respectively.

P18.10.B Natural history of meningiomas - a serial volumetric analysis of 240 tumors

Abstract Background The management of asymptomatic intracranial meningiomas is controversial. Through the assessment of growth predictors, we aimed to create the basis for practicable clinical pathways for the management of these tumors. Material and Methods We volumetrically analyzed meningiomas radiologically diagnosed at our institution between 2003 and 2015. For this purpose, we used exclusively thin-layered MR images (i.e. ≤ 2mm slice thickness). The primary endpoint was tumor growth defined as a 14.35% increase in tumor volume. We identified predictive clinical and radiological characteristics and used the significant variables from a multivariable regression model to construct a decision tree based on the exhaustive Chi-squared Automatic Interaction Detection (exhaustive CHAID) algorithm. Results Of 240 meningiomas, 159 (66.3%) demonstrated growth during a mean observation period of 46.9 months. On multivariable logistic regression analysis, older age (OR=0.979 (0.958-1.000), p=0.048) and presence of calcification (OR=0.442 (0.224-0.872), p=0.019) had a negative predictive value for tumor growth, while T2-signal iso-/hyperintensity (OR=4.415 (2.056-9.479), p&amp;lt;0.001) had a positive predictive value. A decision tree model yielded three growth risk groups based on T2-signal intensity and presence of calcifications with a proportion of growing tumors of 34.1% in the low risk group, 60.0% in the intermediate risk group and 80.2% in the high risk group. Median tumor volume doubling time (Td) was 185.7 months in the low risk, 100.1 months in the intermediate risk and 51.7 months in the high risk group (p&amp;lt;0.001). While 0% of meningiomas in the low and intermediate risk group had a Td of ≤12 months, 8.9% in the high risk group did so (p=0.021). Conclusion Most meningiomas demonstrated growth during follow-up. The presence or absence of calcifications and the signal intensity on T2-weighted imaging allow a practical and simple stratification of meningiomas into low, intermediate and high risk tumors. Small tumors in the low or intermediate risk categories can be monitored with longer follow-up intervals, whereas in the high risk category proactive management decisions can be justified.

Machine learning models for estimating preliminary factory construction cost: case study in Southern Vietnam

Construction of industrial enterprises has become more necessary in recent years. It is critical for project managers to estimate the entire cost of a building project at this early stage. Existing approaches that use operator experience as a mathematical formula. Initial estimates are inaccurate due to the lack of available data points, which leads to overruns in project costs. This research utilizes different machine learning techniques to predict preliminary factory construction cost. Five popular numeric predictive techniques: support vector machine (SVM), artificial neural network (ANN), generalized linear regression (GENLIN), classification and regression-based techniques (CART), exhaustive chi-squared automatic interaction detection (CHAID) are used for baseline and ensemble models. A deep learning neural network (DLNN) is also utilized in this study. The machine learning model is trained and tested on actual data gathered in the southern part of Vietnam. Deep learning outperforms all other machine learning algorithms in this comparison, while the ensemble model of artificial neural networks and generalised linear regression also fared well. Cost estimators can quickly pick the best model for projecting the cost of constructing a preliminary factory by having access to a variety of estimate methodologies.

Predicting body weight of Kalahari Red goats from linear body measurements using data mining algorithms.

Background and Aim:The Kalahari Red goat breed is the finest meat-producing species in South Africa, and its coat color ranges from light to dark red-brown. A practical approach to estimating their body weight (BW) using linear body measurements is still scarce. Therefore, this study aimed to determine the best data mining technique among classification and regression trees (CART), Chi-square automatic interaction detection (CHAID), and exhaustive CHAID (Ex-CHAID) for predicting the BW of Kalahari Red goats.Materials and Methods:This study included 50 Kalahari Red goats (does = 42 and bucks = 8) aged 3–5 years. Body length (BL), heart girth (HG), rump height (RH), height at withers (WH), sex, and age were the essential indicators to estimate BW. The best model was chosen based on the goodness of fit, such as adjusted coefficient of determination (Adj. R2), coefficient of determination (R2), root mean square error (RMSE), standard deviation ratio (SD ratio), mean absolute percentage error, Akaike information criteria, relative approximation error, and coefficient of variation.Results:The SD values of the ratio ranged from 0.32 (CART) to 0.40 (Ex-CHAID). The greatest R2 (%) was established for CART (89.23), followed by CHAID (81.99), and the lowest was established for Ex-CHAID (81.70). CART was established as the preferred algorithm with BL, HG, and WH as critical predictors. The heaviest BW (73.50 kg) was established in four goats with BL higher than 92.5 cm.Conclusion:This study reveals that CART is the optimum model with BL, HG, and WH as the essential linear body measurements for estimating BW for Kalahari Red goats. The updated records will assist the rural farmers in making precise judgments for various objectives, such as marketing, breeding, feeding, and veterinary services in remote areas where weighing scales are unavailable.

Penerapan Metode Klasifikasi Chi-Square Automatic Interaction Detection dan Exhaustive Chi-Square Automatic Interaction Detection

The Chi-Square Automatic Interaction Detection (CHAID) and Exhaustive CHAID methods are nonparametric statistical methods that can be used to classify. CHAID and Exhaustive CHAID were used to determine the significant relationship between the dependent variable and the independent variables based on the chi-square independence test. This study was applied to data on the study period of students of FMIPA UNMUL batch 2014. Based on the CHAID and Exhaustive CHIAD methods, it can be seen that the dependent variable of the study period has a significant relationship with the independent variable, namely the study program and GPA predicate. Where students who graduated on time for the Statistics, Biology and Chemistry study program with a satisfactory GPA predicate of 82 students and with a very satisfactory GPA predicate and cum laude with 46 students. Meanwhile, students who did not graduate on time for the Statistics, Biology and Chemistry study program with an adequate GPA predicate of 5 students, a satisfactory GPA predicate of 41 students, very satisfactory and cum laude with 3 students. Students who graduated on time for the Physics study program were 13 students and those who did not graduate on time were 34 students. The chi-square independence test performed on the CHAID method uses fewer possible categorical pairs than the Exhaustive CHAID method which uses all possible categorical pairs so that it requires a long computational and calculation time.

Study of Hospitalization Costs in Patients with Cerebral Ischemia Based on E-CHAID Algorithm.

Background The aging of the population has led to a rapid increase in the prevalence of most neurological diseases between 1990 and 2016, with a growth rate of up to 117%, which has put enormous pressure on medical insurance funds. As one of the core diseases of disease diagnosis grouping, the hospitalization cost composition and grouping research of patients with cerebral ischemic disease can help to determine scientific payment standards and reduce the economic burden of patients. Aim We aimed to understand the cost composition and influencing factors of hospitalized patients with cerebral ischemic diseases and to identify a reasonable cost grouping scheme. Methods The data come from the homepage of medical records of inpatients with cerebral ischemia in a tertiary hospital in Sichuan Province from 2018 to 2020. After cleaning the data, a total of 5,204 pieces of data were obtained. Nonparametric tests and gamma regression models were used to explore the influencing factors of hospitalization costs. Taking the influencing factors as the predictor variables and the hospitalization cost as the target variable, the exhaustive Chi-squared automatic interaction detector (E-CHAID) algorithm was used to form the costs grouping, and the payment standard of the hospitalization cost for each group was determined. The rationality of cost grouping was evaluated by coefficient of variation (CV) and Kruskal–Wallis H test. Results From 2018 to 2020, the average hospital stay of 5,204 inpatients with cerebral ischemic disease was 10.70 days, and the average hospitalization cost was 17,206.09 RMB yuan. Among the hospitalization costs, diagnosis costs and drug costs accounted for the highest proportion, accounting for 41.18% and 22.38%, respectively, in 2020. Gender, age, admission route, comorbidities and complications, super length of stay (>30 days), and discharge mode had significant effects on hospitalization costs (P < 0.05). Patients were divided into 10 cost groups, and the grouping nodes included comorbidities and complications, discharge mode, age, gender, and admission route. The CV of 9 of the 10 cost groups is less than or equal to 1. The Kruskal–Wallis H test showed that the difference between groups was statistically significant (P < 0.05). Conclusion The cost grouping of patients with cerebral ischemic diseases based on the E-CHAID algorithm is reasonable. This study examined the effects of super length of stay (>30 days), comorbidities and complications, and age on hospitalization cost in patients with cerebral ischemic disease. This study can provide a theoretical basis for advancing the China Healthcare Security Diagnosis Related Groups (CHS-DRG) grouping program and medical expense payment, thereby reducing the disease burden of patients.