In order to achieve the objective, data pertaining to 171 Murrah buffaloes, progeny of 49 sires maintained at Buffalo Research Centre, Hisar were considered for the prediction of lifetime traits from early lactation traits over a period of 24 years from 1990 to 2013. The prediction of phenotypic values of lifetime performance traits viz. Life time milk yield (LTMY), Production life (PL), Milk yield per day of productive life (MY/PL), herd life (HL) and milk yield per day of herd life (MY/HL) were obtained. The equation having the combination of first five lactation traits is recommended for the prediction of phenotypic value of lifetime traits in Murrah buffalo. Accuracy of direct selection for LTMY, PL, MY/PL, HL and MY/HL was 0.42, 0.51, 0.33, 0.51 and 0.54, respectively. The index involving independent variables of first lactation early performance traits had an accuracy of 0.40, 0.39, 0.37, 0.19 and 0.37 for LTMY, PL, MY/PL, HL and MY/HL, respectively. Generally the first lactation milk yield, first peak milk yield and first calving intervals of all lactations contributed maximum to R2 value for all lifetime traits. Therefore, it may be concluded that selection on the basis of first lactation milk yield, first peak yield and first calving interval would also improve the lifetime performance.&nbsp;

Prediction of phenotypic value of lifetime performance traits on the basis of early performance traits in Murrah buffaloes

<div>Background: Many in vivo and laboratory methods have been used to evaluate ruminant feeds. The objective of this experiment was to determine feed intake, digestibility, kinetics of gas production, in vitro organic matter digestibility (OMDv) and metabolizable energy (ME) of Trifolium alexandrinum (T. alexandrinum) and Medicago sativa (M. sativa) at different maturity stages through in vivo and in vitro methods to elaborate predictive equations from chemical constituents.&nbsp; 
 
Methods: In vivo assay was carried out with two groups of five rams male kipped in metabolism cage. Samples of forage, refusal and feces were collected and processed for chemical analysis. In vitro gaz production technique was performed on forage samples. 
 
Result: The bud and early bloom stage of M. sativa recorded the highest digestibility values of the chemical component (P&lt;0.05). The OMDv of M. sativa was 60.65 and 68.26% (P&lt;0.01) for early bloom stage and bud stage, respectively. Crude protein digestibility (CPD) was positively correlated with crude protein rate (R2=0.83, P&lt;0.05). The gas production from the insoluble fraction &ldquo;b&rdquo; fraction showed a positive correlation with acid detergent fiber (ADF) (R2=0.999, P&lt;0.001). However, cumulative gas production at 24h and 48 h of incubation were negatively correlated (P&lt;0.05) with the ADF rate (R2= -0.98 and -0.97, respectively).&nbsp;</div>

<div>Relationship between In vivo, In vitro Parameters and Chemical Composition to Predict the Nutritive Value of Some Legume Forages</div>

<div style="text-align:justify">Background: The soil nematodes can affect the crops in various ways. The plant-parasitic nematodes can lead to severe yield losses. The extent of crop yield loss depends on the susceptibility of the variety or host tolerance, population density of the nematode and various environmental variables. However, no tool is available for the prediction of nematode population buildup in soil therefore it has been difficult to issue advisories for timely management of these pathogens. Here we developed a method to accurately predict the nematode population buildup in soil for its timely management.&nbsp; 
 
Methods: Nematode population index of a plant-parasitic nematode Tylenchorynchus was taken from two crops i.e. mung bean and crotalaria. The model was developed considering various weather variables to predict the population of the Tylenchorynchus in the fields of mung bean and crotalaria. Weather parameters such as maximum and minimum temperature, relative humidity, wind speed and sunshine hours were considered for developing the model for Tylenchorynchus population prediction. Stepwise regression method was applied to predict the nematode population.&nbsp; 
 
Result: The regression analysis between estimated and observed values of Tylenchorynchus population gave the R2 value as 0.98 for mung bean and 0.87 for crotalaria. Well timed prediction can help the growers to apply the required management practices to make it beneficial economically. This method can be extended to predict the population buildup of other serious nematode pests of crops.</div>

<div>Prediction of Nematode Population Dynamics using Weather Variables in Leguminous Crops</div>

Background: Accurate global crop yield predictions are crucial for ensuring food security, effective agricultural planning and climate adaptation. However, existing machine learning and deep learning approaches lack crop-specific feature learning, uncertainty quantification and multiscale spatial contexts, which limits their application in precision agriculture. 
 
Methods: This study presents a hybrid ensemble deep learning framework integrated with a crop-aware transformer encoder with heteroscedastic uncertainty for global multi-crop yield prediction (Maize, rice, wheat, soybean) at a 5-arcminute (~9 km) global resolution. The architecture integrates three key innovations: (1) conservative feature engineering using historical yields (4-year sequences with a 3-year temporal gap), geographic coordinates, climate zone indicators alongside temporal trends and stress indicators; (2) crop-aware multi-head attention different mechanisms with crop-specific Query/Key/Value projections enabling differential pattern learning per crop and (3) heteroscedastic output heads predicting both mean yield (&mu;) and uncertainty (&sigma;) via negative log-likelihood optimization. We implemented rigorous validation using temporal splitting (training: years &le;2013; test: years &gt;2013) and geographic blocking (5-fold GroupKFold with 0.5&deg; spatial blocks) to prevent both spatial and temporal data leakage. 
 
Result: Evaluated on the GlobalCropYield5min dataset (1982-2015) across four crops with 60,000 samples in 34 years. The proposed model achieved an overall test R2 of 0.9281, RMSE of 0.585 t/ha and MAE of 0.379 t/ha, with statistical significance confirmed by a paired t-test (p=0.011). Five-fold geographic cross-validation yielded R2&nbsp;of 0.9337&plusmn;0.0074 and rice R&sup2;=0.9462 (best), maize R&sup2;=0.9268, wheat R&sup2;=0.9201 and soybean R&sup2;=0.8298. Uncertainty quantification achieved excellent calibration (Expected calibration error = 0.024), with empirical coverage matching theoretical values (68% intervals: 69.1% coverage; 95% intervals: 94.8% coverage). Regional analysis showed consistent performance across continents (R&sup2;=0.871-0.941), with data-scarce regions showing the expected performance reduction. Ablation studies confirmed that crop-aware attention contributed +3.4% to R2, multiscale spatial features contributed 5.8% and temporal sequence features contributed +3.66%.

A Multi-crop Ensemble Framework for Yield Prediction

<div>Background: Chickpea is the third major pulse produced globally, with 11.6 million tonnes produced per annum (Merga and Haji, 2019).&nbsp;Sugar alcohols, inulin, starch are all prebiotic carbohydrates found in chickpeas (Johnson et al., 2020).&nbsp;Near-Infrared (NIR) spectroscopy is a non-destructive, versatile and powerful analytical technique. 
&nbsp;&nbsp; 
Methods: Spectral data obtained from NIR spectroscopy requires application of various techniques to extract useful information from spectral data which is further used for building various models for prediction of physical or chemical components presents in agricultural crops. The main aim of this study is to apply various machine learning algorithms especially effective in predicting sugar concentration in chickpea. Sugar prediction models are developed using Linear Regression (LR), Artificial Neural Network (ANN), Random Forest (RF), Support Vector Regression (SVR) and Decision Tree Regression (DTR) algorithms. Performance of the models is evaluated using measures namely, Root Mean Square Error (RMSE), Residual Standard Error (RSE), Coefficient of Determination (R2) and Adjusted Coefficient of Determination (adjusted R2). 
 
Result: It was observed that, RF outperformed all other models in terms of accuracy for predicting sugar component from preprocessed spectra, with RMSE, RSE, R2 and adjusted R2 values of 0.054, 0.062, 0.954 and 0.937, respectively. The accuracy of the ANN model is similar to that of the RF, with minor differences in RMSE, RSE, R2 and adjusted R2 , values of 0.057, 0.067, 0.952 and 0.935.</div>

Comparing Various Machine Learning Algorithms for Sugar Prediction in Chickpea using Near-infrared Spectroscopy

<div style="text-align:justify">Background: Allocation of resources only to a few promising segregating populations that are likely to result in high frequency of transgressive recombinant inbred lines (RILs) for use as pure-line cultivars would help enhance efficiency of breeding self-pollinated crops including dolichos bean. The use of an objective criterion to identify promising segregating populations is therefore assumes importance. The prediction of frequency of transgressive RILs that could be derived from advanced generations of segregating populations of crosses is one such criterion.&nbsp; 
 
Methods: We predicted the frequency of RILs that transgressed the better parent (HA 5) from two reciprocal crosses derived from two elite but genetically diverse parents (HA 4 and HA 5) for four quantitative traits based on estimates of mid parental value [m], additive genetic effects [a] and additive genetic variance [&sigma;2A].&nbsp; 
 
Result: The frequency of transgressive RILs predicted from H 5 &times; HA 4 was higher than those predicted from H 4 &times; HA 5 for primary branches plant-1, pods plant-1 and grain weight plant-1, while it was comparable between the crosses for pod weight plant-1. The required minimum population size was relatively smaller to recover the transgressive RILs from the cross which was predicted to result in higher frequency of RILs than that was predicted to result in lower frequency of RILs.</div>

Predicting the Frequency of Transgressive RILs and Minimum Population Size Required to Recover Them in Dolichos Bean [Lablab purpureus (L.) Sweet]

Background: In agriculture, Crop Yield Prediction (CYP) by Deep Learning (DL) models such as Deep Neural Network (DNN) and Convolutional Neural Network (CNN) is an emerging technique that utilizes weather, soil and previous year yield information for different crops. On the other hand, these models often struggle to effectively handle time-series data, which can lead to challenges in capturing temporal relationships between weather, soil and crop yield factors. This can affect the accuracy of predicting crop yields. 
 
Methods:&nbsp;To address this issue, the Fixed-Adaptive Temporal Attention Network (FAT-AttNet) is proposed in this study for CYP using a time-series crop yield dataset, which includes historical information on soil, weather and crop yields. A normalization technique is used to preprocess the dataset, removing irregularities and empty values. Next, the FAT-AttNet model is trained to predict crop yields. This model consists of fixed and adaptive temporal modules. In the fixed temporal module, 18 parallel one-dimensional (1D) CNNs are used to process the stable factors, such as soil data, to capture temporal features. In the adaptive temporal module, 16 parallel two-dimensional (2D) CNNs capture the dynamic relationships among weather, soil and crop growth factors over time. The results from these models are combined using the lateral connection and an attention strategy is applied to adjust the weight of each factor for more accurate predictions. Furthermore, a Fully Connected (FC) layer and an output layer are used to predict crop yields. 
 
Result: The experiments are conducted using an agricultural crop yield in Indian States dataset (ACY-ISD) to evaluate the performance of the FAT-AttNet model for five major crops in Tamil Nadu using the historical data from 2002 to 2023.&nbsp; The results show that the FAT-AttNet model achieves a Mean Absolute Error (MAE) of 0.05, 0.059, 0.061, 0.058 and 0.074 for groundnut, maize, moong, rice and urad yield predictions, respectively. &nbsp;The Root Mean Square Error (RMSE) for the groundnut, maize, moong, rice and urad yield predictions is 0.217, 0.268, 0.245, 0.235 and 0.266, respectively. The correlation coefficients for groundnut, maize, moong, rice and urad yield predictions are 0.862, 0.846, 0.864, 0.860 and 0.859, respectively. The results of FAT-AttCNN models outperform the Improved Optimizer Function with Long Short-Term Memory (IOF-LSTM), CNN-Bidirectional LSTM (BiLSTM), MayFly Algorithm empowered attention-BiLSTM (MFA-BiLSTM) and Spatio-Temporal Attention-based CNN (STACNN) models. 
&nbsp;

Fixed-adaptive Temporal Attention Network for Predicting Crop Yield

<div style="text-align: justify;">Background: Kashmiri (Kashir) goat is nondescript and highly variable, medium sized, goat genetic resource of Kashmir. The goat possessing excellent survivability and is distributed in Kashmir valley. In Jammu and Kashmir, Gujjars and Bakerwals have adopted the sheep and goat rearing as their primary occupation. Body weights at different ages of life cycle of animals are important traits owing to their direct relation with farmers income. 
Methods: Data on body weight and body measurements viz ear length (EL), body length (BL), body height at wither (BH), chest girth (CG) and tail length (TL) of 162 adult Kasmiri (Kashir) nondescript goats, irrespective of gender, were collected from Febuary 2019 to January 2020 to estimate the body weight from body measurements. The data were subjected to standard statistical analysis using SPSS software. Descriptive statistics and Pearson correlation was worked out to observe association among traits under study. The linear regression was utilized to predict body weight from linear body measurements. 
Result: The average body weight (BW), ear length (EL), body length (BL), body height at wither (BH), chest girth (CG) and tail length (TL) of 36.94&plusmn;0.99 kg, 13.13&plusmn;0.36 cm, 55.48&plusmn;0.43 cm, 65.93&plusmn;0.62 cm, 69.61&plusmn;0.73 cm and 12.94&plusmn;0.15 cm, respectively were observed for Kashmiri goat in the present study. The body weight was having positive and significant correlations with BL, HW, CG and TL. The correlation of body weight with EL and TL was low. Similarly the correlation of body weight with HW was moderate whereas with BL and CG was significantly high indicating these two traits can define body weight of animal more accurately. Based on R2 (coefficient of determination) criteria, the best regression model for predicting body weight of Kashmiri goat was obtained by using combination of BL (body length) and CG (chest girth) with R2 (coefficient of determination) of 0.795. Prediction equation with R2 value of 0.806 was obtained when all traits were included in regression model where as prediction equation with R2 value of 0.589 was obtained when only chest girth was used as independent variable.</div>

Prediction of Body Weight from Linear Body Measurements in Kashmiri (Kashir) Goat

Background: This research paper proposes a new, efficient way of honey quality evaluation through the application of machine learning methods, more specifically on the deep neural network (DNN). Honey is an extremely popular natural sweetener that is not only appreciated for its pleasant taste but also for its numerous health advantages, such as its antioxidant, anti-inflammatory and antimicrobial properties. However, the high popularity of honey has caused its adulteration, particularly through the incorporation of low-cost sugar syrup, to spread rapidly. Such an issue offers critical health hazards to consumers, as honey adulteration cannot be effectively checked through traditional techniques. Therefore, to overcome this challenge, this study aims to present an efficient, smart way to effectively identify honey adulteration. 
 
Methods: This study focuses on predicting the target variable using tabular physicochemical features through a deep learning-based approach. These factors include the colour score, density, water content, pH value, electric conductivity, level of fructose, level of glucose and honey viscosity. All these factors take into consideration the overall description of honey&rsquo;s physical-chemical property. A regression-based model of a Deep Neural Network is trained on a large set of data that comprises more than 200,000 data points. 
 
Result: The performance of the model is excellent and has an R2&nbsp;value of 0.99, showing a level of accuracy that is extremely high to predict the purity level of honey. Notably, the discussion brings out clearly that the model performs well without any signs of either overfitting or underfitting and is very reliable and robust to be used in real applications. The discussion clearly brings out the effectiveness and potential role of Machine Learning and deep learning algorithms used in the context of food quality and food safety. The proposed approach is accurate and efficient and has a high potential to be used to check honey adulteration by various organizations and people. 
&nbsp;

Deep Learning-based Assessment of Honey Purity using Predictive Modelling

<div>Background: Artificial intelligence (AI) is transforming all spheres of life and it has the potential to revolutionize animal husbandry as well. In this regard, an attempt was made to compare two AI techniques for predicting 12-month body weights of animals; viz. Principal Component regression (PCR) and Ordinary Least Squares (OLS) for datasets of Corriedale sheep spanning 11 years.&nbsp; 
 
Methods:&nbsp;PCR models were trained by varying proportions of training and testing datasets. The dataset was subject to PCR before analysis and tested (PCA dataset). A separate dataset was also created by feature selection of the PCA (PCA+FS dataset) variables. 
 
Result:&nbsp;The highest correlation coefficients between test and predicted variables for two datasets (PCA dataset and PCA+FS dataset) created among the multiple models trained using PCR were 0.982 and 0.9741. In terms of error, R2 or correlation coefficient, the PCA dataset performed better than the PCA+FS dataset. The second principal component had the highest explained variance in OLS (86.16%) and the highest coefficient of determination (R2) using OLS was obtained for the PCA dataset viz.&nbsp;0.980. It is concluded that both the algorithms tested in this study were satisfactorily trained in their prediction of the body weights with OLS performing better than PCA in terms of R2 value.</div>