Predictive Discriminant Analysis Research Articles

Emotion regulation from a virtue perspective

BackgroundThe ability to regulate one’s emotional state is an important predictor of several behaviors such as reframing a challenging situation to reduce anger or anxiety, concealing visible signs of sadness or fear, or focusing on reasons to feel happy or calm. This capacity is referred to as emotion regulation. Deficits in this ability can adversely affect one’s adaptive coping, thus are associated with a variety of other psychopathological symptoms, including but not limited to depression, borderline personality disorder, substance use disorders, eating disorders, and somatoform disorders.MethodsThe present study examined emotion regulation in relation to the virtue-based psychosocial adaptation model (V-PAM). 595 participants were clustered based on their Difficulties in Emotion Regulation Scale (DERS) score, producing two clusters (i.e., high functioning vs. low functioning). Then, emotion regulation group membership was discriminated by using five V-PAM virtue constructs, including courage, integrity, practical wisdom, committed action, and emotional transcendence.ResultsResults show that five virtues contribute to differentiating group membership. Practical wisdom was the strongest contributor, followed by integrity, emotional transcendence, committed action, and courage. Predictive discriminant analysis was conducted and 71% of cases were correctly classified. A discussion of the relationship between emotion regulation and virtues was elaborated.ConclusionThe concept of virtue holds significant importance in the comprehension of an individual’s capacity to regulate their emotions, meriting future study.

Differential responses of crop yields to multi-timescale drought in mainland China: Spatiotemporal patterns and climate drivers

Increasingly frequent and severe droughts pose a growing threat to food security in China. However, our understanding of how different crops respond to multi-timescale drought under varying climatic conditions remains limited, hindering effective drought risk management. To address this knowledge gap, we applied spatial principal component analysis (SPCA) to unveil spatiotemporal patterns in annual yields of major grain crops (rice, wheat, maize) and cotton in response to multi-timescale drought, as indicated by the standardized precipitation evapotranspiration index (SPEI) across China. Subsequently, predictive discriminant analysis (PDA) was employed to identify the primary climatic factors driving these response patterns. The findings indicated that drought-induced interannual variability of crop yields were spatially and temporally heterogeneous, closely tied to the timescale used for drought assessment. Crop types displayed distinct responses to drought, evident in the variations of months and corresponding timescales for their strongest reactions. The initial three principal components, capturing over 65 % of drought-related yield variance, unveiled short- to medium-term patterns for rice, maize, and cotton, and long-term patterns for wheat. Specifically, rice was highly susceptible to drought on a 4-month timescale in September, wheat on a 6-month timescale in May, maize on a 3-month timescale in August, and cotton on a 3-month timescale in September. Moreover, the first three discriminant functions explaining over 90 % of the total variance, effectively distinguish spatiotemporal crop yield response patterns to drought. These patterns primarily stem from seasonal climatic averages, with water balance (precipitation minus potential evapotranspiration) and temperature being the most influential variables (p < 0.05). Interestingly, we observed a weak correlation between drought severity and crop yield in humid conditions, with responses tending to manifest over longer timescales. These findings enhance our comprehension of how drought timescales impact crop yields in China, providing valuable insights for the implementation of rational irrigation management strategies.

Soil burrow characterization by 3D image analysis: Prediction of macroinvertebrate groups from biopore size distribution parameters

Although scientific literature focuses mainly on earthworms, all soil macroinvertebrates play an important role in modifying the architecture of the soil pore space and, in turn, in soil functions. Notwithstanding the fact that non-invasive technology, such as X-ray tomography, has long been used to differentiate non-biopores from biopores, it is still difficult to distinguish the specific contribution different macroinvertebrates make to the soil biopore system. Unlike the object-based image analysis approach, when applied to a soil pore system, mathematical morphology permits the user to obtain a very accurate pore size distribution consistent with the physical principle of water retention. The aim of this work was to evaluate the potential of the parameters of this kind of biopore size distribution to differentiate between the burrows of five different macroinvertebrate groups, namely Earthworms, Millipedes juliform, Centipedes, Campodeiform larvae and Elateriform larvae, inoculated into repacked soil mesocosms and incubated (14 days) in the field from where the soil animals were originally collected.A two-fold approach was proposed in this work so as to obtain parameters by both pore size population distributions and Weibull modelling of the cumulative distributions. Then a predictive discriminant analysis was performed on selected parameters by using macroinvertebrate groups as grouping variables and a very good prediction was obtained in both cases. The most useful parameters were the skewness and FFT indices in the first case and the shape parameter α of the Weibull model along with its RMSE in the second one. In addition, topological characterization was performed on gallery-shaped biopores. Vertical deviation was the only parameter that was independent of the individual body size and showed the statistically significant lowest value for the earthworms. The experiment and analyses performed in this work to explore the connection between macroinvertebrate groups and the corresponding biopore size distributions may represent a suitable methodological approach to performing a general investigation into the relationships between soil management and its impact on the system of soil macropores.

Responses of Winter Wheat Yield to Drought in the North China Plain: Spatial–Temporal Patterns and Climatic Drivers

Understanding the winter wheat yield responses to drought are the keys to minimizing drought-related winter wheat yield losses under climate change. The research goal of our study is to explore the response patterns of winter wheat yield to drought in the North China Plain (NCP) and then further to study which climatic factors drive the response patterns. For this purpose, winter wheat yield was simulated by the Environmental Policy Integrated Climate (EPIC) crop model. Drought was quantified by standardized precipitation evapotranspiration index (SPEI), and the contributions of the various climatic factors were evaluated using predictive discriminant analysis (PDA) method. The results showed that the responses of winter wheat yield to different time-scale droughts have obvious spatial differences from the north part to the south part in the NCP. Winter wheat yield is more sensitive to the medium (6–9 months) and long (9–12 months) time-scale droughts that occurred in the key growth periods (April and May). The different response patterns of winter wheat yield to the different time-scale droughts are mainly controlled by temperature and water balance (precipitation minus potential evapotranspiration) in winter in the NCP. Compared with the water balance, temperature plays a more important role in driving the response pattern characteristics. These findings can provide a reference on how to reduce drought influences on winter wheat yield in the NCP.

Identifying Cranberry Juice Consumers with Predictive OPLS-DA Models of Plasma Metabolome and Validation of Cranberry Juice Intake Biomarkers in a Double-Blinded, Randomized, Placebo-Controlled, Cross-Over Study.

Methods to verify cranberry juice consumption are lacking. Predictive multivariate models built upon validated biomarkers may help to verify human consumption of a food using a nutrimetabolomics approach. A 21-day double-blinded, randomized, placebo-controlled, cross-over study was conducted among healthy young women aged 1829. Plasma was collected at baseline and after 3-day and 21-day consumption of cranberry or placebo juice. Plasma metabolome was analyzed using UHPLC coupled with high resolution mass spectrometry. 18 discriminant metabolites in positive mode and 18 discriminant metabolites in negative mode from a previous 3-day open-label study were re-discovered in the present blinded study. Predictive orthogonal partial least squares discriminant analysis (OPLS-DA) models were able to identify cranberry juice consumers over a placebo juice group with 96.9% correction rates after 3-day consumption in both positive and negative mode. This present study revealed 84 and 109 additional discriminant metabolites in positive and negative mode, respectively. Twelve of them were tentatively identified. Cranberry juice consumers were classified with high correction rates using predictive OPLS-DA models built upon validated plasma biomarkers. Additional biomarkers were tentatively identified. These OPLS-DA models and biomarkers provided an objective approach to verify participant compliance in future clinical trials.

Linking tree-ring growth and satellite-derived gross primary growth in multiple forest biomes. Temporal-scale matters

This study links tree-ring growth and gross primary production for a variety of forest types under different environmental conditions across Spain. NOAA-AVHRR satellite imagery data were combined with dendrochronological records and climate data at a fine spatial resolution (1.21 km2) to analyze the interannual variability of tree-ring growth and vegetation activity for different forest biomes from 1981 to 2015. Specifically, we assessed the links between tree-ring width indices (TRWi), the Normalized Difference Vegetation Index (NDVI) and a variety of environmental conditions, represented by climatic variables (air temperature, precipitation, evapotranspiration and water balance) and elevation. The impact of these variables on tree growth was assessed by means of the Predictive Discriminant Analysis (PDA). Results reveal a general positive and significant relationship between inter-annual variability of the NDVI at a high spatial resolution (1.21 km2) and tree-ring growth. Maximum correlations between NDVI and tree-ring growth were recorded when cumulative NDVI values were considered, in some cases covering long time periods (6–10 months), suggesting that tree growth is mainly related to Gross Primary Production (GPP) at annual scale. The relationship between tree-ring growth and inter-annual variability of the NDVI, however, strongly varies between forest types and environmental conditions.

A landscape of overlapping risks for wolf-human conflict in Wisconsin, USA

Managing risk requires an adequate understanding of risk-factors that influence the likelihood of a particular event occurring in time and space. Risk maps can be valuable tools for natural resource managers, allowing them to better understand spatial characteristics of risk. Risk maps can also support risk-avoidance efforts by identifying which areas are relatively riskier than others. However, risks, such as human-carnivore conflict, can be diverse, multi-faceted, and overlapping in space. Yet, efforts to describe risk typically focus on only one aspect of risk. We examined wolf complaints investigated in Wisconsin, USA for the period of 1999–2011. We described the spatial patterns of four types of wolf-human conflict: livestock depredation, depredation on hunting hounds, depredation on non-hound dogs, and human health and safety concerns (HHSC). Using predictive landscape models and discriminant functions analysis, we visualized the landscape of risk as a continuous surface of overlapping risks. Each type of conflict had its own unique landscape signature; however, the probability of any type of conflict increased closer to the center of wolf pack territories and with increased forest cover. Hunting hound depredations tended to occur in areas considered to be highly suitable wolf habitat, while livestock depredations occurred more regularly in marginal wolf habitat. HHSC and non-hound dog depredations were less predictable spatially but tended to occur in areas with low housing density adjacent to large wildland areas. Similar to other research evaluating the risk of human-carnivore conflict, our data suggests that human-carnivore conflict is most likely to occur where humans or human property and large carnivores co-occur. However, identifying areas of co-occurrence is only marginally valuable from a conservation standpoint and could be described using spatially-explicit human and carnivore data without complex analytical approaches. These results challenge our traditional understanding of risk and the standard approach used in describing risk. We suggest that a more comprehensive understanding of the risk of human-carnivore conflict can be achieved by examining the spatial and non-spatial factors influencing risk within areas of co-occurrence and by describing the landscape of risk as a continuous surface of multiple overlapping risks.

Predicting outcome of matches using pressure index: evidence from Twenty20 cricket

In this study, an attempt is made to predict the outcome of Twenty20 cricket matches applying predictive discriminant analysis on pressure index. The classification functions of discriminant analysis are developed based on a training sample of 118 international Twenty20 matches and its validity has been tested on a holdout sample of another 73 matches. The cross-validation exercise yields a maximum of 90.7 percent and 78.4 percent correct classification in the training sample and holdout sample respectively. Finally, the findings of the proposed method are compared with other target rescheduling methods used for the purpose of match prediction.

Admission Criteria for Educational Leadership Doctoral Students in One U.S. Doctoral Program

Aim/Purpose: The purpose of this study was to explore relationships between preadmission criteria and doctoral student performance ratings and to develop a model to predict student persistence in one doctoral program of educational leadership. Background: Individuals responsible for program admission decisions have a responsibility to minimize bias in the candidate selection process. Despite an interest in doctoral degree completion, few researchers have examined preadmission criteria and the ability to predict doctoral student performance, particularly in education programs. Methodology: Preadmission variables and postacceptance performance ratings were used in this cross-sectional predictive study (Type 5; Johnson, 2001) of 102 doctoral students in one educational leadership program. Analyses included descriptive statistics, a Pearson r correlation matrix, and predictive discriminant analysis. Contribution: In addition to strengthening the extant literature base, we attempted to respond to the charge levied by other researchers for faculty members in educational preparation programs to reassess current practices used to recruit and retain students. Findings: Using predictive discriminant analysis, we determined that separate models for students of color and White students most accurately predicted program performance, indicating that a one-size fits all approach was not optimal. The GRE-Q and undergraduate GPA were useful predictors of doctoral student persistence. Additionally, the GRE-V and graduate GPA were also useful predictors but differentially so for students of color and White students. Recommendations for Practitioners: We found value in using the GPA and GRE in admission decisions with some modifications. Programs directors are advised to evaluate their own selection processes to understand the utility of their preadmission criteria. Recommendation for Researchers: Although the functions that worked best in predicting continuance were grouped by ethnicity in this study for our students, future researchers might consider disaggregation by gender or some other characteristic to optimally identify a model specific to the student groups represented in their sample. Impact on Society: Working from an activist stance, we use our awareness of the positive correlation between degree attainment and socio-economic mobility in the United States, coupled with the existing realities of students of color who seek access to a space within the dominant culture, to urge admission committees to evaluate closely the variables used in their admission selection and to understand to what extent the selection process results in a fair selection across student groups. Future Research: Future studies could be conducted to understand why these differences exist. Other variables for future researchers to consider include time since the candidates obtained their master’s and bachelor’s degrees, the length of time to obtain those degrees, and the type of degree obtained.

Partial Least Squares Discriminant Analysis and Bayesian Networks for Metabolomic Prediction of Childhood Asthma.

To explore novel methods for the analysis of metabolomics data, we compared the ability of Partial Least Squares Discriminant Analysis (PLS-DA) and Bayesian networks (BN) to build predictive plasma metabolite models of age three asthma status in 411 three year olds (n = 59 cases and 352 controls) from the Vitamin D Antenatal Asthma Reduction Trial (VDAART) study. The standard PLS-DA approach had impressive accuracy for the prediction of age three asthma with an Area Under the Curve Convex Hull (AUCCH) of 81%. However, a permutation test indicated the possibility of overfitting. In contrast, a predictive Bayesian network including 42 metabolites had a significantly higher AUCCH of 92.1% (p for difference < 0.001), with no evidence that this accuracy was due to overfitting. Both models provided biologically informative insights into asthma; in particular, a role for dysregulated arginine metabolism and several exogenous metabolites that deserve further investigation as potential causative agents. As the BN model outperformed the PLS-DA model in both accuracy and decreased risk of overfitting, it may therefore represent a viable alternative to typical analytical approaches for the investigation of metabolomics data.

Real Time Face Recognition Based on Face Descriptor and Its Application

This paper presents a real time face recognition based on face descriptor and its application for door locking. The face descriptor is represented by both local and global information. The local information, which is the dominant frequency content of sub-face, is extracted by zoned discrete cosine transforms (DCT). While the global information, which also is the dominant frequency content and shape information of the whole face, is extracted by DCT and by Hu-moment. Therefore, face descriptor has rich information about a face image which tends to provide good performance for real time face recognition. To decrease the dimensional size of face descriptor, the predictive linear discriminant analysis (PDLDA) is employed and the face classification is done by kNN. The experimental results show that the proposed real time face recognition provides good performances which indicated by 98.30%, 21.99%, and 1.8% of accuracy, FPR, and FNR respectively. In addition, it also needs short computational time (1 second).

A New Approach for Improving Classification Accuracy in Predictive Discriminant Analysis

The focus of a predictive discriminant analysis is to improve classification accuracy, and to obtain statistically optimal classification accuracy or hit rate is still a challenge due to the inherent variability of most real life dataset. Improving classification accuracy is usually achieved with best subset of relevant predictors obtained by using classical variable selection methods. The goal of variable selection methods is to choose the best subset (or training sample) of relevant variables that typically reduces the complexity of a model and makes it easier to interpret, improves the classification accuracy of the model and reduces the training time. However, a statistically optimal hit rate can be achieved if the training sample meets a near optimal condition by resolving any significant differences in the variances for the groups formed by the dependent variable. This paper proposes a new approach for obtaining a near optimal training sample that will produce a statistically optimal hit rate using a modified winsorization with graphical diagnostic. In application to real life data sets, the proposed new approach was able to identify and remove legitimate contaminants in one or more predictors in the training sample, thereby resolving any significant differences in the variances for the groups formed by the dependent variable. The graphical diagnostic associated with the new approach, however, provides a useful visual tool which served as an alternative graphical test for homogeneity of variances.

Reducing the Need for Postsecondary Remediation Using Self-efficacy to Identify Underprepared African-American and Hispanic Adolescents

Introduction. Postsecondary remedial education is a major problem for Higher Education affecting retention rates, degree completion rates and cost of enrollment. Identification of students who are underprepared for managing the rigors of college and who would benefit from early intervention is necessary to reduce the need for postsecondary remediation. This study examines student perceptions of self-efficacy and use of self-regulatory behaviors as a possible indicator for need for remediation.Method. Sixty-three 10th and 11th grade Texas students (52.5% African American, 44.3% Hispanic, 3.2% others) participated in the study. Two self-regulation inventories, the SRI-HS and LASSI-HS, were administered to determine student self-efficacy and their use of self-regulatory behaviors. A Predictive Discriminant Analysis was used to successfully classify level of self-efficacy of a group of African American and Hispanic adolescents into 3 mutually exclusive groups, [high (81%), moderate (65%) or low (72%)] based on scores from the two self-regulation inventories.Results. A Kruskal-Wallis analysis revealed areas of the SRI-HS and LASSI-HS inventories where low self-efficacy African American and Hispanic students scored significantly lower than their high self-efficacy counterparts on self-reported use of self-regulatory skills and behaviors. Both inventories revealed problems in managing distractions, attitude, concentration, help-seeking, study strategies, goal setting, attention, selecting the main idea, and test strategies; important skills needed by under-prepared students and would be useful in intervention. Attitude was the lowest percentile score (15th percentile) for high and low self-efficacy groups. The PDA model correctly identified students (72.9% of the grouped cases) making such a model useful for predicting students with low academic self-efficacy (72%) to identify underprepared students; students that could benefit from self-efficacy and self-regulation intervention.Discussion. The results support the hypothesis that students that have high self-efficacy tend to use more self-regulatory skills and behaviors than students with lower self-efficacy. This finding suggests high self-efficacy is a link to higher achievement and therefore, less likelihood of needing post-secondary remediation. Suggestions for improving student self-efficacy self-regulation and attitude in both high and low self-efficacy, underprepared students are offered to decrease the likelihood of need for post-secondary remediation.

Discriminant analysis for prediction and classification of farmers based on adoption of drought coping mechanisms

Discriminant analysis for prediction and classification of farmers based on adoption of drought coping mechanisms

Early Detection of Progressive Adolescent Idiopathic Scoliosis: A Severity Index.

Early detection of progressive adolescent idiopathic scoliosis (AIS) was assessed based on 3D quantification of the deformity. Based on 3D quantitative description of scoliosis curves, the aim is to assess a specific phenotype that could be an early detectable severity index for progressive AIS. Early detection of progressive scoliosis is important for adapted treatment to limit progression. However, progression risk assessment is mainly based on the follow up, waiting for signs of rapid progression that generally occur during the growth peak. Sixty-five mild scoliosis (16 boys, 49 girls, Cobb Angle between 10 and 20°) with a Risser between 0 and 2 were followed from their first examination until a decision was made by the clinician, either considering the spine as stable at the end of growth (26 patients) or planning to brace because of progression (39 patients). Calibrated biplanar x-rays were performed and 3D reconstructions of the spine allowed calculating six local parameters related to main curve deformity. For progressive curve 3D phenotype assessment, data were compared with those previously assessed for 30 severe scoliosis (Cobb Angle > 35°), 17 scoliosis before brace (Cobb Angle > 29°) and 53 spines of nonscoliosis subjects. A predictive discriminant analysis was performed to assess similarity of mild scoliosis curves either to those of scoliosis or nonscoliosis spines, yielding a severity index (S-index). S-index value at first examination was compared with clinical outcome. At the first exam, 53 out of 65 predictions (82%) were in agreement with actual clinical outcome. Approximately, 89% of the curves that were predicted as progressive proved accurate. Although still requiring large scale validation, results are promising for early detection of progressive curves. 2.

A predictive model of adolescent persistence in counseling

This archival study examined personal and environmental characteristics associated with premature termination from, or persistence in, counseling among a sample of adolescents (n=74) who presented for services at a community counseling clinic. The predictive discriminant analysis yielded a model with a 74.6% hit rate for counseling outcome. Internalizing behaviors and time spent on the wait list accounted for the most variance in counseling persistence.

Strategies to predict and improve eating quality of cooked beef using carcass and meat composition traits in Angus cattle.

Product quality is a high priority for the beef industry because of its importance as a major driver of consumer demand for beef and the ability of the industry to improve it. A 2-prong approach based on implementation of a genetic program to improve eating quality and a system to communicate eating quality and increase the probability that consumers' eating quality expectations are met is outlined. The objectives of this study were 1) to identify the best carcass and meat composition traits to be used in a selection program to improve eating quality and 2) to develop a relatively small number of classes that reflect real and perceptible differences in eating quality that can be communicated to consumers and identify a subset of carcass and meat composition traits with the highest predictive accuracy across all eating quality classes. Carcass traits, meat composition, including Warner-Bratzler shear force (WBSF), intramuscular fat content (IMFC), trained sensory panel scores, and mineral composition traits of 1,666 Angus cattle were used in this study. Three eating quality indexes, EATQ1, EATQ2, and EATQ3, were generated by using different weights for the sensory traits (emphasis on tenderness, flavor, and juiciness, respectively). The best model for predicting eating quality explained 37%, 9%, and 19% of the variability of EATQ1, EATQ2, and EATQ3, and 2 traits, WBSF and IMFC, accounted for most of the variability explained by the best models. EATQ1 combines tenderness, juiciness, and flavor assessed by trained panels with 0.60, 0.15, and 0.25 weights, best describes North American consumers, and has a moderate heritability (0.18 ± 0.06). A selection index (I= -0.5[WBSF] + 0.3[IMFC]) based on phenotypic and genetic variances and covariances can be used to improve eating quality as a correlated trait. The 3 indexes (EATQ1, EATQ2, and EATQ3) were used to generate 3 equal (33.3%) low, medium, and high eating quality classes, and linear combinations of traits that best predict class membership were estimated using a predictive discriminant analysis. The best predictive model to classify new observations into low, medium, and high eating quality classes defined by the EATQ1 index included WBSF, IMFC, HCW, and marbling score and resulted in a total error rate of 47.06%, much lower than the 60.74% error rate when the prediction of class membership was based on the USDA grading system. The 2 best predictors were WBSF and IMFC, and they accounted for 97.2% of the variability explained by the best model.

Empirically Exploring Higher Education Cultures of Assessment

Although touted as beneficial to student learning, cultures of assessment have not been examined adequately using validated instruments. Using data collected from a stratified, random sample (N = 370) of U.S. institutional research and assessment directors, the models tested in this study provide empirical support for the value of using the Administrators’ Survey of Assessment Cultures as an assessment tool. The resulting first order model provided good fit statistics and included five factors: a) Faculty Perceptions, b) Use of Data, c) Sharing, d) Compliance or Fear Motivators, and e) Normative Purposes for Assessment. Internal consistency estimates were also good (Cronbach’s α = .792 to .922). Use of the Schmid-Leiman solution afforded further support for the higher order factor, Culture of Assessment. Predictive discriminant analysis correctly classified the majority of the respondents (80.7%) as belonging to a student learning or accreditation group based on the linear classification scores. Implications for assessment leadership, practice, and future research are offered.

The Relationship Between Continuous Identity Disturbances, Negative Mood, and Suicidal Ideation.

To examine the relationship between continuous identity and a measure of depression, anxiety, and stress as well as suicidal ideation using 2 validated measures of continuous identity. A total of 246 subjects recruited from the Amazon Mechanical Turk subject pool who completed a full survey in November 2014 were included in the analyses. Stress, anxiety, and depression severity were measured using the Depression, Anxiety, and Stress Scale. Continuous identity was measured with the Venn continuous identity task and the me/not me continuous identity task. Multiple regression analyses revealed continuous identity disturbances were significantly associated with depressed mood (R (2) = 0.37, P < .01). Continuous identity also predicted suicide severity, even after controlling for demographic factors, negative life events, and depressed mood. Additionally, predictive discriminant analysis revealed continuous identity, depression severity, and negative life events correctly classified 74.1% of participants into high and low suicide risk groups. Lack of continuous identity predicted both depression and suicidality severity. Integration of perceived identities may be a worthwhile goal for behavioral interventions aimed at reducing depressed mood and suicidality.

Significance of Discriminant Analysis in Prediction of Pregnancy in IVF Treatment

Abstract Many factors play an important role in prediction of infertility treatment outcome (for example, female age and quality of oocytes or embryos are the most important prognostic factors concerning positive IVF outcome). The purpose of this study was to identify a set of variables that could fulfill criteria for prediction of pregnancy in IVF patients through the application of data mining – using the discriminant analysis method. The principle of this method is to establish a set of rules that allows one to place multi-dimensional objects into one of two analyzed groups (pregnant or not pregnant). Six hundred and ten IVF cycles were included in the analysis and the following variables were taken into consideration: female age, number and quality of retrieved oocytes, number and quality of embryos, number of transferred embryos, and outcome of treatment. Discriminant analysis allowed for the creation of a model with a 51.22% correctness of prediction to achieve pregnancy during IVF treatment and with 74.07% correctly predicted failure of pregnancy. Therefore, the created model is more suitable for the prediction of a negative outcome (lack of pregnancy) during IVF treatment and offers an option for adjustments to be made during infertility treatment.