Discriminant Function Techniques Research Articles

Utilizing Selection Indices and Discriminant Function Analysis to Enhance Seed Yield in Sesame (Sesamum indicum L.)

The current study focuses on harnessing genetic variation and employing selection indices for evaluating thirty-four advanced sesame breeding lines, alongside four checks (DS-5, DSS-9, JTS-8, and TKG-22). The evaluation encompassed the assessment of fifteen quantitative traits during the summer of 2022 at the AICRP on Sesame and Niger, MARS, UAS, Dharwad. Particularly noteworthy was the high genotypic and phenotypic coefficient variation observed for traits such as the number of secondary branches per plant, yield per plant (g) and seed yield (Kg/ha). The highest heritability coupled with the greatest genetic advance over the mean was detected for the number of primary branches per plant, suggesting a predominant role of additive genetic components in their expression and indicating a promising avenue for direct selection. The study involved the construction of thirty-one selection indices using the discriminant function technique, which incorporated five key traits: seed yield per plant (g) (X1), days to maturity (X2), number of productive capsules per plant (X3), thousand seed weight (g) (X4), and oil content (%) (X5). Among these various selection indices, the one comprising all component characters (X1, X2, X3, X4, and X5) exhibited the highest expected genetic advance and relative efficiency.

Selection indices for grain yield improvement in durum wheat [Triticum turgidum L. subsp. durum (Desf.) Husn.

The plant breeder has certain desired plant characteristics in his mind while selecting for particular genotypes and for this he/she applies various weight to different traits for arriving at decisions. This suggests the use of a selection index that gives proper weight to each of the two or more characters to be considered. Among forty genotypes of durum wheat thirty-one selection indices, involving grain yield per plant and four yield contributing components, were constructed using the discriminant function technique. The superiority of selection based on index increases with an increase in the number of characters under selection. The selection index constructed based on five characters viz., grain yield per plant, days to maturity, plant height, biological yield per plant and harvest index exhibited maximum relative efficiency and selection efficiency. The plant breeder might be interested in maximum genetic gain with minimum number of characters. The maximum relative efficiency per character was observed in selection index comprised of days to maturity and plant height (X3 + X4) followed by plant height and biological yield per plant (X3 + X4 ).

Discriminate function analysis in cowpea (Vigna unguiculata (L.) Walp.)

The discriminant function technique was used to construct selection indices in forty two genotypes of cowpea (Vigna unguiculata (L.) Walp.). Sixty three selection indices involving seed yield per plant and its five components were constructed using the discriminant function technique. The efficiency of selection increased considerably when the selection was based on two character combinations i.e. the number of pods per plant and the number of seeds per pod followed by an index based on three characters viz., the number of clusters per plant and the number of pods per plant and seed yield per plant. The use of these indices is advocated for selecting high yielding genotypes of cowpea as in practice the plant breeder might be interested in the maximum gain with the minimum number of characters.

Selection indices for yield improvement in bread wheat (Triticum aestivum L.)

The discriminant-function technique was used to construct selection indices in 48 genotypes of bread wheat (Triticum aestivum L.). Sixty-three selection indices involving grain yield per plant and its five components were constructed using the discriminant function technique. The efficiency of selection increased with the inclusion of more characters in the selection index. The index based on four characters viz., grain yield per plant, the number of grains per main spike, grain weight per main spike and biological yield per plant recorded the highest genetic gain and relative efficiency. The use of these indices is advocated for selecting high yielding genotypes of bread wheat.

Selection indices in groundnut (Arachis hypogaea L.)

Fifty diverse genotypes of groundnut were evaluated in a randomized block design with three replications for the study of selection indices during summer 2014. Thirty-one selection indices involving pod yield per plant (X1) and four yield components viz., kernel yield per plant (X2) harvest index (X3), number of mature pods per plant (X4) and biological yield per plant (X5) were constructed using the discriminant function technique. Discriminant function analysis indicated that selection efficiency of the function was improved by increasing the number of characters in the index. Among the single character indices, biological yield per plant exhibited higher genetic advance and relative efficiency over straight selection for pod yield per plant. The index based on five characters viz., pod yield per plant, kernel yield per plant, harvest index, number of mature pods per plant and biological yield per plant recorded the highest genetic advance as well as relative efficiency and selection efficiency. These characters could be advantageously exploited in the groundnut breeding programmes.

Selection indices in Bread Wheat [Triticum aestivum L.

The discriminant-function technique was used to construct selection indices in 52 genotypes of bread wheat (Triticum aestivum L.). Sixty-three selection indices involving grain yield per plant and its five components were constructed using discriminant function technique. In general, the more the number of characters included in a selection index, the better was its performance. The index based on five characters viz., grain yield per plant, 100 grain weight, days to maturity, harvest index and number of effective tillers per plant which had highest genetic advance and relative efficiency of 155.77g and 1867.41%, respectively followed by an index based onfive characters i.e. grain yield per plant, biological yield per plant, 100-grain weight, days to maturity, harvest index which possessed genetic gain and relative efficiency of 53.85g and 1856.78% respectively. The use of both these indices is advocated for selecting high yielding genotypes ofbread wheat.

Selection indices for yield improvement in bread wheat under late sown condition

Investigation was carried-out to assess the selection indices in 40 genotypes of wheat (Triticum aestivum L.) in limited irrigation under late sowing condition. Sixty-three selection indices, involving grain yield per plant and five yield components, were constructed using the discriminant function technique. The efficiency of selection increased with the inclusion of more number of characters in the index. The selection index based on six characters viz., grain yield per plant, days to maturity, number of productive tillers per plant, number of grain per main spike, biological yield per plant and harvest index under limited irrigated condition exhibited maximum gain and relative efficiency. It is expected that grain yield could be improved if due consideration is given to these traits in future breeding programme of wheat.

Sex determination from the foramen magnum

Abstract This paper presents the findings of a study to investigate the value of measurements of the foramen magnum in relation to the determination of biological sex. The study performed a large-scale analysis of 250 adult individuals (♂=144, ♀=106) from a modern Swiss population utilising CT data acquired in the context of the Virtopsy ® Project. Data was analysed using discriminant function and binary logistic regression techniques and visual observations of the morphology of the foramen magnum were made to contextualise the findings. The study found that with all variables, males and females were equally classified at 66% accuracy by discriminant function analysis, where binary logistic regression was comparatively poor at classifying females despite an overall classification rate of 66.4%. The morphology of the foramen magnum was classified in 88% of the sample by visual assessment according to seven shape types; however 12% of the sample was unclassified due to shape irregularities arising from asymmetry and rim complexities. This study concludes that while measurements of the foramen magnum appear to demonstrate statistically significant differences between the sexes, the ability to discriminate sex may be limited by morphological variation. This may also influence the accuracy to which sex can be determined in individual crania and thus the isolated use of this method is not advisable unless as a suggestive finding when other features of assessment are absent or limited.

The Effects of Initially Misclassified Data on the Effectiveness of Discriminant Function Analysis and Finite Mixture Modeling

Classification procedures are common and useful in behavioral, educational, social, and managerial research. Supervised classification techniques such as discriminant function analysis assume training data are perfectly classified when estimating parameters or classifying. In contrast, unsupervised classification techniques such as finite mixture models (FMM) do not require, or even use if available, knowledge of group status to estimate parameters or classifying. This study investigates the impact of two types of misclassification errors on the classification accuracy of discriminant function analysis (both linear [LDA] and quadratic [QDA]) and FMM for two groups with a single predictor. Analytic and Monte Carlo results are provided for a variety of misclassification scenarios to investigate the performance of the two methods. Discriminant function techniques recovered the highest overall percentages of correctly classified data, whereas FMM captured higher percentages of the smaller group when group sizes are unequal. LDA marginally outperformed QDA under misclassified conditions.

Monitoring of viral cancer progression using FTIR microscopy: A comparative study of intact cells and tissues

Fourier transform infrared microspectroscopy (FTIR-MSP) is an analytical method with a promising potential for detecting the spectral changes due to cancerous changes in cells. The purpose of the present study is monitoring biochemical spectral changes accompanying viral cancer progression in cells and tissues using FTIR-MSP. As a model system, we used cells in culture which were transformed to malignant cells by infection with murine sarcoma virus (MuSV) and cervical tissues at different neoplastic stages. In order to devise a systematic follow-up of the cancer progression, it was essential first to determine and validate consistent and significant spectral biomarkers, which can evidently discriminate between normal and cancerous cells/tissues. Then these biomarkers were used for the characterization and classification of early stages of malignant transformation utilizing discriminant classification function techniques. Our study points out that malignancy progression can be eminently graded for both cell lines and tissues. For example, using the array of four biomarkers: A 2958 A 2852 + A 2923 , A 1121 / A 1015 , A 1171 / A 1152 and | A 1082 − A 1056 | A 1028 , we attained that the classification accuracies of different premalignant stages of cell lines and tissues were varied between 89.5 and 97.4%. These results strongly support the potential of developing FTIR microspectroscopy as a simple, reagent free method for early detection and accurate differentiation of premalignant stages.

Early spectral changes of cellular malignant transformation using Fourier transform infrared microspectroscopy

Fourier transform infrared microspectroscopy (FTIR-MSP) is potentially a powerful analytical method for identifying the spectral properties of biological activity in cells. The goal of the present research is the implementation of FTIR-MSP to study early spectral changes accompanying malignant transformation of cells. As a model system, cells in culture are infected by the murine sarcoma virus (MuSV), which induces malignant transformation. The spectral measurements are taken at various postinfection time intervals. To follow up systematically the progress of the spectral changes at early stages of cell transformation, it is essential first to determine and validate consistent and significant spectral parameters (biomarkers), which can evidently discriminate between normal and cancerous cells. Early stages of cell transformation are classified by an array of spectral biomarkers utilizing cluster analysis and discriminant classification function techniques. The classifications indicate that the first spectral changes are detectable much earlier than the first morphological signs of cell transformation. Our results point out that the first spectral signs of malignant transformation are observed on the first and third day of postinfection (PI) (for NIH/3T3 and MEF cell cultures, respectively), while the first visible morphological alterations are observed only on the third and seventh day, respectively. These results strongly support the potential of developing FTIR microspectroscopy as a simple, reagent-free method for early detection of malignancy.

A neural computational model of ICSI outcomes from sperm source and other features

Objective: With the success and increasing popularity of ICSI, accurately predicting outcomes from this artificial reproductive technique becomes highly desirable. We examined a large dataset of ICSI outcomes and associated clinical features such as sperm source, which technician performed the procedure, number of ova retrieved and maternal age, and modeled outcomes using neural computational techniques. Neural computation employs computer programs that implement mathematical models based loosely on the function of networks of biological neurons. Statistical feature extraction, allowing examination of the significance of individual features such as sperm source, was employed using Wilk’s Generalized Likelihood Ratio Test (GLRT). Design: Neural computational modeling of retrospective data. Materials/Methods: We modeled fertilization outcomes from ICSI from technician (7 different technicians performed the procedure,) sperm source (testis extraction, MESA, EEJ,) number of ova retrieved and maternal age using neUROn2++, a set of C++ programs we developed to implement neurocomputational algorithms, discriminant function and regression techniques using the Cygwin (Red Hat) GNU C++ port for Windows (Microsoft) and Visual C++ (Microsoft) distributed across Pentium (Intel) platforms. Wilk’s GLRT was implemented in neUROn2++ for statistical feature extraction. A dataset of 1070 unique ICSI outcomes was randomized into a training set of 1349 exemplars and 449 test exemplars. N1/N2 cross-validation was employed. Results: Overall model accuracy was ROC AUC 0.922 (sensitivity 86.9%, specificity 79.0%) in the training set, and 0.897 (sensitivity 82.8%, specificity 75.1%) in the test set. Feature extraction using Wilk’s GLRT in reverse regression is shown in the table. All features were significant to the model (all p <1e-5), however, with p <1e-106, technician was highly significant.TableRegression analysis of neural computational model to predict ICSI outcomes.Featurep-valueTechnician1.04e-107No. ova retrieved9.84e-45Testis extraction3.81e-21Maternal age7.70e-20MESA2.33e-10EEJ1.52e-06p-values indicate significance of the individual feature to the likelihood of fertilization with ICSI. Open table in a new tab p-values indicate significance of the individual feature to the likelihood of fertilization with ICSI. Conclusions: We modeled ICSI outcome with high accuracy (ROC AUC 0.897 in the test set) using a neural computational approach. In order to determine the significance of individual features to the model’s outcome, regression based on Wilk’s GLRT was performed. All features examined (sperm source, number of ova retrieved, maternal age and which technician performed the procedure) were statistically significant to the model’s outcome. Supported by: NIH P01 HD36289 to Dolores J Lamb, Larry I Lipshultz and Craig Niederberger.

Projection-slice synthetic discriminant functions for optical pattern recognition

The projection-slice synthetic discriminant function (PSDF) filter is introduced and proposed for distortion-invariant pattern-recognition applications. The projection-slice theorem, often used in tomographic applications for medical imaging, is utilized to implement a distortion-invariant filter. Taking M projections from one training image and combining them with (N - 1)M projections taken from another N - 1 training image accomplishes this. With the projection-slice theorem, each set of these M projections can be represented as M one-dimensional slices of the two-dimensional Fourier transform of the particular training image. Therefore, the PSDF filter has the advantage of matching each of the training images with at least M slices of their respective Fourier transforms. This filter is theoretically analyzed, numerically simulated, and experimentally implemented and tested to verify the simulation results. These tests show that the PSDF filter significantly outperforms the matched-filter and the basic synthetic discriminant function technique for the particular images used.

The application of methods of non-linear dynamics for the improved and predictive recognition of patients threatened by sudden cardiac death

This study introduces new methods of non-linear dynamics (NLD) and compares these with traditional methods of heart rate variability (HRV) and high resolution ECG (HRECG) analysis in order to improve the reliability of high risk stratification. Simultaneous 30 min high resolution ECG's and long-term ECG's were recorded from 26 cardiac patients after myocardial infarction (MI). They were divided into two groups depending upon the electrical risk, a low risk group (group 2, n = 10) and a high risk group (group 3, n = 16). The control group consisted of 35 healthy persons (group 1). From these electrocardiograms we extracted standard measures in time and frequency domain as well as measures from the new non-linear methods of symbolic dynamics and renormalized entropy. Applying discriminant function techniques on HRV analysis the parameters of non-linear dynamics led to an acceptable differentiation between healthy persons and high risk patients of 96%. The time domain and frequency domain parameters were successful in less than 90%. The combination of parameters from all domains and a stepwise discriminant function separated these groups completely (100%). Use of this discriminant function classified three patients with apparently low (no) risk into the same cluster as high risk patients. The combination of the HRECG and HRV analysis showed the same individual clustering but increased the positive value of separation. The methods of NLD describe complex rhythm fluctuations and separate structures of non-linear behavior in the heart rate time series more successfully than classical methods of time and frequency domains. This leads to an improved discrimination between a normal (healthy persons) and an abnormal (high risk patients) type of heart beat generation. Some patients with an unknown risk exhibit similar patterns to high risk patients and this suggests a hidden high risk. The methods of symbolic dynamics and renormalized entropy were particularly useful measures for classifying the dynamics of HRV.

Identification of change structure in statistical process control

In order to diagnose properly quality problems that occur in manufacturing the diagnostician, be it human or computer, must be privy to various sources of information about the process and its behaviour. This paper describes how neural networks and Bayesian discriminant function techniques can be used to provide knowledge of how a product characteristic changed, i.e. shift in mean or variability, when so noted by the control chart application. Such information is useful because there usually exists some underyling knowledge about the physical phenomena in question that relates the behaviour of the observed characteristic to its processing variables. When a change in the process is detected by the appropriate statistical method, a feature vector of process-related statistics is used to identify the change structure as a shift in mean or variance. This paper addresses various issues concerned with this problem, namely: process change detection, feature vector selection, training patterns, and error rates. S...

Primary appraisal of the Strange Situation: A cross-cultural analysis of preseparation episodes.

Studies in various countries–among them Germany, Holland, Israel, Japan, Sweden, and the United States–have reported Strange Situation distributions that differ markedly across and within cultures, thus raising doubts as to whether infant behavior in the Strange Situation can be regarded as a valid index of the security of attachment, at least in a cross-cultural context. It is proposed here that a fuller understanding of infant behavior in the Strange Situation requires an assessment of what Connell and Goldsmith (1982) have referred to as the infant's primary appraisal of this procedure. A crossnational comparison of data from seven laboratories ( N = 498) was carried out in order to determine (a) whether the preseparation episodes made any difference in attachment classifications and (b) whether infant behavior in different countries was indeed the same before separation from the mother. Furthermore, procedural variations have been taken into account. Using multivariate discriminant function techniques, information obtained during preseparation episodes was found to discriminate between attachment classifications. At the same time, however, in a series of multivariate analyses of variance, a relatively small number of cross-cultural differences in preseparation behavior were found. With the exception of the Israeli kibbutz and Swedish infants, our data suggest that infants in different countries make similar primary appraisals of the Strange Situation.

Differences in accident characteristics among elderly drivers and between elderly and middle-aged drivers

A total of 5,853 elderly drivers, aged 55 and over, and 8,210 middle-aged drivers 36–50 years old, all of whom were involved in accidents during 1986, were compared using discriminant function and univariate techniques. Substantial differences were found in accident characteristics not only between middle-aged and elderly drivers but also between three different age groups of the elderly. Support was found for the suspicion that, while the elderly have fewer accidents than younger drivers, their safety record, based on driving exposure, with noncrash convictions as a surrogate measure of kilometres driven, is worse, and, in addition, they are more often at fault in accident involvements.

Feature extraction from spectral and other data by the principal components and discriminant function techniques

Principal component analysis is applied to the interpretation of 13C-n.m.r. spectra and to the resolution of mass spectral data. A procedure is given for determining the relative amounts of pure components, with and without the use of pure mass lines, in mass spectra of mixtures. The use of the Fisher discriminant method in combination with the principal components technique is demonstrated in the treatment of trace element data on hair for environmental purposes. The importance of feature generation and selection is emphasized.

Nest-Site Selection by the Red-Tailed Tropicbird

Abstract Red-tailed Tropicbirds (Phaethon rubricauda) nest under shrubs in loosely associated groups on Christmas Island, Central Pacific Ocean. Aggressive interactions among adults at nest sites are high; yet, large areas of superficially similar habitat contain no tropicbird nests at all. We compared physiognomic features of nest sites with those of two sets of randomly generated "nest sites" by discriminant function analysis. Of those variables measured, the number of stems within a nest space and the amount of peripheral cover influenced the tropicbirds' choice of nest sites. Differences among localities existed. Areas where tropicbirds nested had shrubs with few stems and more peripheral cover. Preferred nest sites were rare in areas other than those in which tropicbirds were nesting already. This study demonstrates the utility of the discriminant function technique in quantifying nest-site selection and nest-site availability when the appropriate null data sets are used.

The use of discriminant function techniques in reliability assessment and data classification

In some cases of reliability assessment and data classification, the influence of loading and environmental factors are not difficult to see. In many instances, however, this is not the case, and particularly with mechanical equipment the relevance of such factors can be almost impossible to define. In this situation the significance of each can often be quantified through the use of statistical discrimination techniques. Discrimination function techniques are described as a means of classifying reliability information in a meaningful form. The techniques and software described offer a method of quantifying the significance of environmental parameters as a measure of their influence to some function, which in most reliability problems will be to one or more modes of failure or success. The process is ideally suited to multi-variable dependent problems and results in a series of weighting parameters which quantify the relevance of each variable to the overall outcome. The techniques are demonstrated on several examples of mechanical seal data and shown to produce discrimination, and variable ranking information useful to both data classification and failure prediction.