Linear Ridge Regression Research Articles

Novel and versatile artificial intelligence algorithms for investigating possible GHSR1α and DRD1 agonists for Alzheimer's disease.

Hippocampal lesions are recognized as the earliest pathological changes in Alzheimer's disease (AD). Recent researches have shown that the co-activation of growth hormone secretagogue receptor 1α (GHSR1α) and dopamine receptor D1 (DRD1) could recover the function of hippocampal synaptic and cognition. We combined traditional virtual screening technology with artificial intelligence models to screen multi-target agonists for target proteins from TCM database and a novel boost Generalized Regression Neural Network (GRNN) model was proposed in this article to improve the poor adjustability of GRNN. R-square was chosen to evaluate the accuracy of these artificial intelligent models. For the GHSR1α agonist dataset, Adaptive Boosting (AdaBoost), Linear Ridge Regression (LRR), Support Vector Machine (SVM), and boost GRNN achieved good results; the R-square of the test set of these models reached 0.900, 0.813, 0.708, and 0.802, respectively. For the DRD1 agonist dataset, Gradient Boosting (GB), Random Forest (RF), SVM, and boost GRNN achieved good results; the R-square of the test set of these models reached 0.839, 0.781, 0.763, and 0.815, respectively. According to these values of R-square, it is obvious that boost GRNN and SVM have better adaptability for different data sets and boost GRNN is more accurate than SVM. To evaluate the reliability of screening results, molecular dynamics (MD) simulation experiments were performed to make sure that candidates were docked well in the protein binding site. By analyzing the results of these artificial intelligent models and MD experiments, we suggest that 2007_17103 and 2007_13380 are the possible dual-target drugs for Alzheimer's disease (AD).

Comparison of linear, penalized linear and machine learning models predicting hospital visit costs from chronic disease in Thailand

Generally, health care costs from chronic diseases have positive skew and this gives problems on using traditional statistical models. Machine learning is a conventional method producing accurate prediction with large sample size. However, much of the comparison performance between statistical methods and machine learning for such data remains scattered. This study aimed to compare linear, penalized linear and machine learning models for their prediction performance of hospital visit costs from chronic disease, in Thailand. A total of 18,342 hospital visit records were obtained from Suratthani tertiary hospital in southern Thailand, which contained data from 2016 on chronic patients of Diagnosis-Related Groups (DRGs). The prediction performance on hospital visit costs by linear, penalized linear and machine learning models were compared using both original dataset and datasets expanded in size two- and four-fold by using bootstrap. The mean age of patients was 56.3 ± 22.6 years with 55.6% of visits by males. The median hospital cost was 16,662 Baht per visit. The random forest (RF) model had the best predictive performance of hospital visit costs for all sizes of dataset with the smallest prediction errors, whereas ridge linear regression had the poorest prediction performance with the largest prediction errors. Machine learning models had better prediction performance with enlarged sample sizes whereas linear and penalized linear models did not. On modeling big data for prediction, machine learning models are preferable, whereas linear and penalized linear models' predictions are not affected by increasing the sample size.

НАУКОВО-МЕТОДИЧНИЙ ПІДХІД ДО ОЦІНЮВАННЯ МУЛЬТИПЛЕКСИВНОЇ ЕФЕКТИВНОСТІ ЕКОЛОГІЧНИХ ПОДАТКІВ У ЗАБЕЗПЕЧЕННІ НАЦІОНАЛЬНОЇ БЕЗПЕКИ

Increased interest in balancing environmental, energy and economic policy issues leads to the actualization of the search for comprehensive instruments of state influence. The intersectoral nature of environmental taxation shows a significant potential for its impact on the simultaneous provision of various areas of national security. The existence of multiplexed transmission links between the individual components of national security necessitated the development of a scientific and methodological approach to a comprehensive assessment of the impact of environmental taxes on national security in the context of building an optimal system of environmental taxation. The research is built as a sequence of interrelated stages. The selection of environmental taxes that have the potential to have multiplex effects on national security was determined using the Granger causality test. The components of national security that depend on the impact of environmental taxes are defined as environmental, energy and economic security. Transmission multiplexes that arise between individual components of national security are taken into account by aggregating them into an integral index according to the Kolmogorov-Gabor method. To determine the impact of individual environmental taxes and their combinations on national security, multiple linear and nonlinear ridge regression equations are constructed. The maximum level of national security that can be achieved as a result of environmental taxation is determined using a multi-parameter optimization model. Solving the optimization model allows determining the optimal structure of environmental taxes in the country. Determining the integrated levels of environmental, energy and economic security that correspond to the optimal level of national security achieved as a result of environmental taxation should be done through spectral analysis, which will take into account the links between the components of national security. The proposed approach creates a sound methodological basis for determining the directions of development of environmental taxes in order to achieve global and national public policy goals.

Risk Assessment of Myocardial Infarction for Diabetics through Multi-Aspects Computing

INTRODUCTION: Myocardial infarction (MI) is a type of cardiovascular disease. Cardiovascular disease is the major side effect of diabetes. It causes damage to heart muscle due to interruption in the blood flow. The chance of getting this disease is high in diabetes patients.OBJECTIVES: To choose a dataset with features related to diabetes, parameters of ECG and risk factors of MI for effective prediction. Predict myocardial infarction in both type-1 and type-2 diabetic patients using regression techniques. Recognise the best algorithm.METHODS: Multiple linear regression, ridge regression and lasso regression are existing techniques in addition to which proposed technique lasso regression is used to develop a model for prediction. The trained models are compared to know better performing algorithm. Estimation statistics namely confidence and prediction intervals are used to show the amount of uncertainty in predicted values. The statistical measures in regression analysis namely root mean squared error and r_squared value are used to evaluate and compare algorithms.RESULTS: The proposed algorithm ‘lasso regression’ has achieved better values of RMSE and r_squared as 0.418 and 0.2278 respectively compared to remaining techniques.CONCLUSION: Best performance of proposed algorithm was noticed and hence using lasso regression for prediction of myocardial infarction in diabetes patients gives better results.

Conditional Generative Adversarial Networks for Individualized Treatment Effect Estimation and Treatment Selection.

Treatment response is heterogeneous. However, the classical methods treat the treatment response as homogeneous and estimate the average treatment effects. The traditional methods are difficult to apply to precision oncology. Artificial intelligence (AI) is a powerful tool for precision oncology. It can accurately estimate the individualized treatment effects and learn optimal treatment choices. Therefore, the AI approach can substantially improve progress and treatment outcomes of patients. One AI approach, conditional generative adversarial nets for inference of individualized treatment effects (GANITE) has been developed. However, GANITE can only deal with binary treatment and does not provide a tool for optimal treatment selection. To overcome these limitations, we modify conditional generative adversarial networks (MCGANs) to allow estimation of individualized effects of any types of treatments including binary, categorical and continuous treatments. We propose to use sparse techniques for selection of biomarkers that predict the best treatment for each patient. Simulations show that MCGANs outperform seven other state-of-the-art methods: linear regression (LR), Bayesian linear ridge regression (BLR), k-Nearest Neighbor (KNN), random forest classification [RF (C)], random forest regression [RF (R)], logistic regression (LogR), and support vector machine (SVM). To illustrate their applications, the proposed MCGANs were applied to 256 patients with newly diagnosed acute myeloid leukemia (AML) who were treated with high dose ara-C (HDAC), Idarubicin (IDA) and both of these two treatments (HDAC+IDA) at M. D. Anderson Cancer Center. Our results showed that MCGAN can more accurately and robustly estimate the individualized treatment effects than other state-of-the art methods. Several biomarkers such as GSK3, BILIRUBIN, SMAC are identified and a total of 30 biomarkers can explain 36.8% of treatment effect variation.

The influence of specific aspects of occupational stress on security guards’ health and work ability: detailed extension of a previous study

In our earlier study of security guards, we showed that higher occupational stress was associated with health impairments (metabolic syndrome, diabetes, hypertension, cardiovascular diseases) and work disability. The aim of this study was to further explore the association of specific occupational stressors with health impairments and work disability parameters in 399 Serbian male security guards (aged 25–65 years). Ridge linear regression analysis revealed that, after controlling for age, body mass index, and smoking status, professional stressors including high demands, strictness, conflict/uncertainty, threat avoidance and underload were significant positive predictors of fasting glucose, triglycerides, total and LDL cholesterol, blood pressure, heart rate, Framingham cardiovascular risk score, and temporary work disability. The security profession is in expansion worldwide, and more studies are needed to establish precise health risk predictors, since such data are generally lacking.

A data science approach for advanced solid polymer electrolyte design

Solid polymer electrolytes (SPEs) stand to revolutionize battery technology innovation by making batteries non-flammable, flexible, and more sustainable. However, SPE breakthroughs are limited by the highly time and resource-intensive nature of battery research. Even when suitable materials are discovered, optimizing the composition and experimental conditions presents another critical barrier to SPE realization. In this work, a data-driven approach to SPE development is presented. First, data is collected and analyzed from published literature, and then supplemented with independent experimentation to complete the SPE dataset. Then, six different models (linear regression, lasso regression, ridge regression, decision tree, random forest, and radial basis function SVM) were tested. The random forest model is identified as the most suitable model with the greatest predictive capability, validated by independent experimentation and by comparing predicted activation energies to those reported in literature using raw predictions from the model. The random forest model is calculated to predict conductivity with a root-mean-square-error of 0.332 log(S/cm). By applying machine learning to incorporate important parameters of SPE synthesis, this study provides a foundation for accelerated SPE innovation.

A Six-Protein Activity Signature Defines Favorable Response to Selinexor Treatment for Patients with Diffuse Large B-Cell Lymphoma (DLBCL)

Introduction: Selinexor is an oral selective inhibitor of XPO1 that was recently approved by the US FDA for treatment of adult patients with relapsed or refractory DLBCL, after at least two lines of systemic therapy. Approval was based on results of the Phase 2b SADAL study, which had an overall response rate of 29% in the primary analysis population (clinicaltrials.gov: NCT02227251). We herein investigated molecular markers of response to selinexor from patients treated on the SADAL study. Methods: Exome sequencing was performed on pre-selinexor treatment biopsies from 55 patients and used to compare mutation frequencies between 21 responder patients (best overall response of complete response [CR], 6; and partial response [PR], 15) and 34 non-responder patients (stable disease [SD], 8; and progressive disease [PD], 26). Additionally, RNA sequencing was performed on a subset of 33 patients and gene expressions were used to infer activities of regulatory proteins with the VIPER algorithm. Differences in inferred protein activities between responder and non-responder patients were assessed using four machine learning algorithms: linear regression (LR), linear discriminant analysis (LDA), ridge regression (RR) and random forest (RF). Model performance was estimated using leave-one-out cross-validation (LOOCV). A separate comparison was performed in the subset of 12 patients with germinal B-cell like (GCB) DLBCL. Results: Our analysis of genes commonly mutated in DLBCL revealed that non-responder patients more frequently harbored mutations in KMT2D (35% non-responders, 14% responders). Examination of the specific types of KMT2D mutations showed that the vast majority were loss-of-function frameshift or nonsense mutations (13 of 15 mutations), indicating they could have functional relevance to disease biology. The activities of 5,742 regulatory proteins were successful inferred from RNA sequencing performed on 33 patients. Unsupervised clustering identified two outlier samples that were removed from further analysis. The remaining 31 patients consisted of 16 responders (CR, 5; and PR, 11) that were compared to 15 non-responders (PD, 15). Dimension reduction of the 5,742 protein activities (filtering proteins with low variation, poor VIPER imputation, and strong linkage) resulted in 680 independent informative regulatory protein activities used for predicting selinexor response. Different numbers of regulatory proteins were iteratively input into the machine learning models to compare responders and non-responders. The best performance model was achieved using only six proteins (ASH1L, ZNF471, RRN3, CD248, ZNF750, INHBA) (Figure 1A), and had an area under the receiver operating characteristic curve (AUC) of 0.917, 0.925, 0.883, and 0.875, for the LDA, LR, RF and RR, models, respectively (p < 0.05, permutation test) (Figure 1B). A final integrated model combining the four methods achieved an AUC = 0.929 (p < 0.05, permutation test, AUC 95% CI: [0.831-1] DeLong non-parametric method) (Figure 1C). Similar results were obtained using 5-fold cross validation with the six-protein activity signature (integrated model AUC = 0.858, AUC 95% CI: [0.72-0.997]). Finally, we focused separately on 12 patients with germinal B-cell like (GCB) DLBCL, (6 responder patients, and 6 non-responder patients). Using LOOCV with the top three protein activities associated with selinexor response in GCB-DLBCL (COL1A1, INHBA, and CNOT2) resulted in remarkably high accuracy, with an integrated model AUC = .972 (p < 0.05, permutation test, AUC 95% CI: [0.895-1]). Discussion and Conclusions: The six proteins used for defining the DLBCL selinexor response signature are not typically associated with a role in DLBCL but have been implicated in cancer biology in other contexts. Notably, INHBA was found as predictor of response in the full set and also the GCB subtype patients, suggesting that activin/inhibin signaling could be important for response to selinexor in patients with DLBCL, especially the GCB subtype. Our results produced a protein activity signature that could be useful for identifying patients with DLBCL likely to respond well to selinexor treatment, which will be validated in a larger independent sample set. Figure 1 Disclosures Walker: Vigeo Therapeutics: Consultancy; Karyopharm: Current Employment, Current equity holder in publicly-traded company. Alvarez:DarwinHealth, Inc: Current Employment, Current equity holder in private company. Chang:Karyopharm Therapeutics Inc: Current Employment. Shah:Karyopharm Therapeutics Inc: Current Employment, Current equity holder in publicly-traded company. Shacham:Karyopharm: Current Employment, Current equity holder in publicly-traded company, Patents & Royalties: (8999996, 9079865, 9714226, PCT/US12/048319, and I574957) on hydrazide containing nuclear transport modulators and uses, and pending patents PCT/US12/048319, 499/2012, PI20102724, and 2012000928) . Califano:DarwinHealth, Inc: Consultancy, Current Employment, Current equity holder in private company, Other: Founder. Landesman:Karyopharm Therapeutics Inc: Current Employment, Current equity holder in publicly-traded company.

Myopia-related stepwise and quadrant retinal microvascular alteration and its correlation with axial length.

To investigate the characteristics of retinal vascular alterations in patients with varying degrees of myopia, and to determine correlated variables and alteration patterns over different retinal zones. A total of 208 right eyes of 208 patients with myopia were enrolled and divided into mild, moderate, high, and extreme myopia groups. The macular vasculature in superficial, deep, and whole retinal layers was imaged with optical coherence tomography angiography (OCTA). The vessel densities over the whole annular zone, four quadrants, and six annuli were quantified as fractal dimension using a customized software. Simple linear regressions and ridge regression were used to determine and compare explanatory variables of microvascular density. The microvascular density within the superficial, deep and whole retinal layers was highest in mild myopia group, second highest in moderate myopia group, second lowest in high myopia group and lowest in extreme myopia group. Within the same layer, differences between any two groups were significant (all p < 0.05). Among the four quadrants, only the inferior nasal (IN) quadrant showed no difference in microvascular density between mild and moderate groups. Ridge regression indicated that microvascular density values in all three layers were more strongly correlated with axial length (AL) (scaled estimates -0.139, -0.103, -0.154; all p < 0.001) than with spherical equivalent (SE) (scaled estimates -0.052, -0.096, -0.057; all p < 0.05). We characterized a stepwise and quadrant alteration of retinal microvascular density from mild to extreme myopia, which was more strongly affected by axial elongation, although both AL and SE were meaningful indicators.

Forecasting industrial aging processes with machine learning methods

Accurately predicting industrial aging processes makes it possible to schedule maintenance events further in advance, ensuring a cost-efficient and reliable operation of the plant. So far, these degradation processes were usually described by mechanistic or simple empirical prediction models. In this paper, we evaluate a wider range of data-driven models, comparing some traditional stateless models (linear and kernel ridge regression, feed-forward neural networks) to more complex recurrent neural networks (echo state networks and LSTMs). We first examine how much historical data is needed to train each of the models on a synthetic dataset with known dynamics. Next, the models are tested on real-world data from a large scale chemical plant. Our results show that recurrent models produce near perfect predictions when trained on larger datasets, and maintain a good performance even when trained on smaller datasets with domain shifts, while the simpler models only performed comparably on the smaller datasets.

Data-driven design of glasses with desirable optical properties using statistical regression

In this study, we used a data-driven approach to build models for assisting the design of new glasses with high refractive index and low dispersion. Our models, which are based on multiple linear regression and kernel ridge regression, achieved high accuracy in predicting optical properties of glasses based on their composition alone. Using the predictions of these models as a guide, we fabricated new glasses in our laboratory. In agreement with model predictions, these glasses had promising optical properties. This work therefore demonstrates a successful example of data-driven materials design and can be used as a template for designing glasses or other materials with other desirable properties.

Exchange Spin Coupling from Gaussian Process Regression

Heisenberg exchange spin coupling between metal centers is essential for describing and understanding the electronic structure of many molecular catalysts, metalloenzymes, and molecular magnets for potential application in information technology. We explore the machine-learnability of exchange spin coupling beyond linear regression, which has not been studied yet. We employ Gaussian process regression, since it can potentially deal with small training sets (as likely associated with the rather complex molecular structures required for exploring spin coupling) and since it provides uncertainty estimates ("error bars") along with predicted values. We compare a range of descriptors and kernels for 257 small dicopper complexes and find that a simple descriptor based on chemical intuition, consisting only of copper-bridge angles and copper-copper distances, clearly outperforms several more sophisticated descriptors when it comes to extrapolating toward larger experimentally relevant complexes. Exchange spin coupling is similarly easy to learn as the polarizability, while learning dipole moments is much harder. The strength of the sophisticated descriptors lies in their ability to linearize structure-property relationships, to the point that a simple linear ridge regression performs just as well as the kernel-based machine-learning model for our small dicopper data set. The superior extrapolation performance of the simple descriptor is unique to exchange spin coupling, reinforcing the crucial role of choosing a suitable descriptor and highlighting the interesting question of the role of chemical intuition vs systematic or automated selection of features for machine learning in chemistry and material science.

LMFAO: An engine for batches of group-by aggregates

LMFAO is an in-memory optimization and execution engine for large batches of group-by aggregates over joins. Such database workloads capture the data-intensive computation of a variety of data science applications. We demonstrate LMFAO for three popular models: ridge linear regression with batch gradient descent, decision trees with CART, and clustering with Rk-means.

The relational data borg is learning

This paper overviews an approach that addresses machine learning over relational data as a database problem. This is justified by two observations. First, the input to the learning task is commonly the result of a feature extraction query over the relational data. Second, the learning task requires the computation of group-by aggregates. This approach has been already investigated for a number of supervised and unsupervised learning tasks, including: ridge linear regression, factorisation machines, support vector machines, decision trees, principal component analysis, and k-means; and also for linear algebra over data matrices. The main message of this work is that the runtime performance of machine learning can be dramatically boosted by a toolbox of techniques that exploit the knowledge of the underlying data. This includes theoretical development on the algebraic, combinatorial, and statistical structure of relational data processing and systems development on code specialisation, low-level computation sharing, and parallelisation. These techniques aim at lowering both the complexity and the constant factors of the learning time. This work is the outcome of extensive collaboration of the author with colleagues from RelationalAI, in particular Mahmoud Abo Khamis, Molham Aref, Hung Ngo, and XuanLong Nguyen, and from the FDB research project, in particular Ahmet Kara, Milos Nikolic, Maximilian Schleich, Amir Shaikhha, Jakub Závodný, and Haozhe Zhang. The author would also like to thank the members of the FDB project for the figures and examples used in this paper. The author is grateful for support from industry: Amazon Web Services, Google, Infor, LogicBlox, Microsoft Azure, RelationalAI; and from the funding agencies EPSRC and ERC. This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 682588.

Learning Models over Relational Data Using Sparse Tensors and Functional Dependencies

Integrated solutions for analytics over relational databases are of great practical importance as they avoid the costly repeated loop data scientists have to deal with on a daily basis: select features from data residing in relational databases using feature extraction queries involving joins, projections, and aggregations; export the training dataset defined by such queries; convert this dataset into the format of an external learning tool; and train the desired model using this tool. These integrated solutions are also a fertile ground of theoretically fundamental and challenging problems at the intersection of relational and statistical data models. This article introduces a unified framework for training and evaluating a class of statistical learning models over relational databases. This class includes ridge linear regression, polynomial regression, factorization machines, and principal component analysis. We show that, by synergizing key tools from database theory such as schema information, query structure, functional dependencies, recent advances in query evaluation algorithms, and from linear algebra such as tensor and matrix operations, one can formulate relational analytics problems and design efficient (query and data) structure-aware algorithms to solve them. This theoretical development informed the design and implementation of the AC/DC system for structure-aware learning. We benchmark the performance of AC/DC against R, MADlib, libFM, and TensorFlow. For typical retail forecasting and advertisement planning applications, AC/DC can learn polynomial regression models and factorization machines with at least the same accuracy as its competitors and up to three orders of magnitude faster than its competitors whenever they do not run out of memory, exceed 24-hour timeout, or encounter internal design limitations.

Forecasting of customer demands for production planning by local [formula omitted]-nearest neighbor models

Demand forecasting is of major importance for manufacturing companies since it provides a basis for production planning. However, demand forecasting can be a difficult task because customer demands often fluctuate due to several influences. Methods of nonlinear dynamics have shown promising results in numerous applications but they have mostly been neglected in the context of demand forecasting. This paper evaluates the forecasting performance of local k-nearest neighbor models, which base on the theory of dynamical systems, in a comprehensive empirical study utilizing a large dataset of industrial time series of the M3-Competition. After a broad literature review, the theoretical background is described. Subsequently, different parameter configurations and model selection strategies are compared. A locally constant mean and a locally constant median are compared to locally linear regression models with four different regularization methods and different parameter configurations. In this comparison, the locally constant mean and the locally linear ridge regression with high regularization parameters provide the best trade-offs between forecast accuracy and computation times. Finally, these models achieve a high performance regarding low forecast errors, short computation times as well as high service levels in an inventory simulation compared to established benchmark methods. In particular, they obtain the best results among all applied methods regarding short time series. Moreover, they achieve the lowest errors considering the original accuracy criterion of the M3-Competition. Hence, local k-nearest neighbor models can be regarded as a valid alternative for demand forecasting in an industrial context, accomplishing high forecast accuracy with short computation times.

A New Robust Epigenetic Model for Forensic Age Prediction.

Forensic DNA phenotyping refers to an emerging field of forensic sciences aimed at the prediction of externally visible characteristics of unknown sample donors directly from biological materials. The aging process significantly affects most of the above characteristics making the development of a reliable method of age prediction very important. Today, the so-called "epigenetic clocks" represent the most accurate models for age prediction. Since they are technically not achievable in a typical forensic laboratory, forensic DNA technology has triggered efforts toward the simplification of these models. The present study aimed to build an epigenetic clock using a set of methylation markers of five different genes in a sample of the Italian population of different ages covering the whole span of adult life. In a sample of 330 subjects, 42 selected markers were analyzed with a machine learning approach for building a prediction model for age prediction. A ridge linear regression model including eight of the proposed markers was identified as the best performing model across a plethora of candidates. This model was tested on an independent sample of 83 subjects providing a median error of 4.5 years. In the present study, an epigenetic model for age prediction was validated in a sample of the Italian population. However, its applicability to advanced ages still represents the main limitation in forensic caseworks.

Machine learning approach to ship fuel consumption: A case of container vessel

An improvement of the marine vessel's fuel consumption will provide efficiency and profitability in ship management since fuel cost is one of the biggest operating cost. However, estimation of the fuel consumption of marine vessels is a difficult issue, because the fuel consumption rate of the vessel is directly dependent on multiple external factors such as the condition of the main engine, cargo weight, ship draft, sea condition, weather condition, etc. Nowadays, statistical models have been established based on actual ship data, and the fuel consumption of the vessel has been tried to be estimated as best as possible. In this study, various prediction models such as Multiple Linear Regression, Ridge and LASSO Regression, Support Vector Regression, Tree-Based Algorithms, Boosting Algorithms have been established for a container ship. The accuracy of the models is determined by the K-fold cross-validation. Error metrics such as root mean squared error, mean absolute error, coefficient of determination are employed in order to evaluate the correctness of estimation models and correlation analysis between variables is accomplished. Parameters such as main engine rpm, main engine cylinder values, scavenge air, shaft indicators are found highly correlated with fuel consumption. Under the influence of various external factors on fuel consumption, the nearest estimation of the actual fuel consumption data is made by multiple linear regression and ridge regression with 0.0001 root mean square error, 0.002 mean absolute error and %99.9 coefficient of determination score.

Modeling and simulation of recurrent phenotypic and genomic selections in plant breeding under the presence of epistasis

Recurrent selection is an important breeding method for population improvement and selecting elite inbreds or fixed lines from the improved germplasm. Recently, a computer simulation tool called QuMARS has been developed, which allows the simulation and optimization of various recurrent selection strategies. Our major objective in this study was to use the QuMARS tool to compare phenotypic recurrent, marker-assisted recurrent, and genomic selections (abbreviated respectively as PS, MARS and GS) for both short- and long- term breeding procedures. For MARS, two marker selection models were considered, i.e., stepwise (Rstep) and forward regressions (Forward). For GS, three prediction models were considered, i.e., genomic best linear unbiased predictors (GBLUP), ridge regression (Ridge), and regression by Moore-Penrose general inverse (InverseMP). To generate genotypes and phenotypes for a given individual during simulation, one additive and two epistasis genetic models were considered with three levels of heritability. Results demonstrated that selection responses from GBLUP-based GS and MARS (Forward) were consistently greater than those from PS under the additive model, particularly in early selection cycles. In contrast, selection response from PS was consistently superior over MARS and GS under epistatic models. For the two epistasis models, total genetic variance and the additive variance component were increased in some cases after selection. Through simulation, we concluded that GS and PS were effective recurrent selection methods for improved breeding of targeted traits controlled by additive and epistatic quantitative trait loci (QTL). QuMARS provides an opportunity for breeders to compare, optimize and integrate new technology into their conventional breeding programs.

Exploiting Ensemble Learning for Rainfall Prediction using Meta Regressors and Meta Classifiers

Intense rainfall produces flooding even on dry soil. As heavy rainfall is one of the causes for flooding it is necessary to predict the Rainfall to take necessary precautions for people who are living in risk zone areas. Prediction of Rainfall tomorrow is done accurately using Machine Learning regression and classification Techniques. For Rainfall prediction multiple attributes like Windspeed, Precipitation, Cloudcover, Humidity, Temperature and RainfallToday are considered to predict Rainfall Tomorrow. An ensemble approach is used where predictions from Regression models such as Linear Regression, Polynomial Regression, Ridge Regression and Lasso regression are stacked together and fed as new attributes to Meta Regressor along with Support Vector Regression for making final predictions. Also, predictions from classifications models such as Gaussian Naive Bayes, K-nearest neighbor, Support vector Machine and Random Forest are stacked together and fed as new attributes to Meta Classifier along with Logistic regression which is a binary classifier for higher predictive performance.