Search

Preharvest crop yield prediction is an important field of agriculture. Timely and precise prediction of crop yield can help the farmers in making right assessment of gains they can expect from their crop and can also help them in prior planning to avoid any crisis. Crop yield is dependent on many controlled and uncontrolled factors. An accurate assessment of these factors in terms of their impact on crop yield can help the farmers in taking steps to increase the produce of their crop. The present study is focused on identifying the influential environmental factors in the study of wheat crop yield prediction. Various feature selection techniques categorized under different feature selection methods: filter methods, wrapper methods and embedded methods were employed to select the most influential factors. The results of feature selection were further validated by using the feature subset generated in random forest prediction model. Among all the feature selection techniques, recursive feature elimination technique under embedded methods was found to be the most efficient feature selection technique. Among the environmental factors, maximum and minimum temperature, relative humidity, evaporation, wind direction and solar radiation were found to be the most prominent and influential factors in relation to wheat crop yield.

Non-communicable diseases are fast assuming significance due to an increase in incidence and related secondary complications. Cardiovascular diseases (CVD) are one of the most significant examples of organ crosstalkinterference. Over time, patients with cardiovascular diseases express markers that lead to dysfunction of other organs. Due to CVD onset, patients suffer from acute or chronic kidney dysfunction, causing bidirectional cardiac and renal interaction leading to cardio-renal diseases. Despite current testing methods, models for personalized CVD diagnosis and its association with renal diseases are inadequate. In this view, we designed a preliminary algorithm to stratify CVD patients’ biomarker expression according to their ethnicity, gender, and age. In this study, we performed a systematic review of the clinical data of CVD patients. A preliminary algorithm for precision diagnosis is developed using baseline clinical factors, history, and CVD proteomic markers and correlated with and without modelling. Pathway analysis is also performed for assessing co-expression of kidney-specific biomarkers with CVD markers to indicate bi-directional interaction of heart and kidney. The results show that the expression of CVD biomarkers is dependent on both the gender and ethnicity of an individual. Algorithm modelling of CVD markers with a patient-specific history and characteristic shows a new set of markers to be considered when diagnosing an individual and predict the onset of other complications. Personalized diagnosis is important to provide tailored treatment as it may prevent multiple organ dysfunction due to organ crosstalk. The base model can be extended to a more extensive algorithm and implemented as a guided treatment and timely decision tool for clinicians to assess the risk of developing CVD and CVD-dependent renal complications in a more personalized way.

Soil is a major and important natural resource, which not only support human life but also furnish commodities for ecological and economic growth. Ecological risk, such as degradation of soil pose a serious threat to the ecosystem. The high-stress level of heavy metals like chromium, copper, cadmium, etc., produce ecological risks, which include: decrease in the fertility of the soil; reduction in crop yield and degradation of metabolism of living beings, and hence ecological health. The ecological risk associated demands the assessment of heavy metal stress levels in soils. As the rate of stress level of heavy metals is exponentially increasing in recent times, it is apparent to assess or predict heavy metal contamination in soil. The assessment will help the concerned authorities to take corrective as well as preventive measures to enhance the ecological, and hence economic growth. This study reviews the efficient assessment models to predict contamination of soil due to heavy metal.

Eyesight is a key sense that human beings possess. Reports shows that color blindness, one form of color vision deficiency(CVD), touches about 8% of the males and 0.5% females (1). Patients of color blindness are unable to differentiate certain colors called confusion colors. Inability to differentiate red results in red blindness, green in green blindness, and blue/yellow in blue-yellow blindness. Routine activities of daily life become a challenge for the colorblind due to this inability. Digital solutions in the area of computer vision can be used to make lives easier for the colorblind (2). The authors propose a fuzzy based method to simulate and compensate confusion colors in images for varying degrees of color blindness. Experiments were conducted on images for simulation and color compensation. Observations of the color compensation was measured with normal vision and colorblind patients. It resulted in 91% accuracy for red green color blindness. The efficiency of simulation was measured by evaluation with state-of-the-art online tools and literature. The results show 95.8% accuracy. The results obtained affirms the efficiency of the proposed fuzzy method.

In today’s scenario, the classification process is required for numerous purposes, including planning of modern cities, agriculture production, and monitoring of environmental factors. So far, the optimal categorization level has remained unattainable. Manual feature-based approaches, unsupervised feature learning methods, and supervised feature learning methods are the three types of remote-sensing images classification methods available. Supervised deep learning methods have recently been widely used in a variety of remote sensing applications, including as object detection and land use scene classification. In contrast to traditional computer vision tasks that only use the spatial context, this article can improve hyperspectral image classification by utilizing both spatial context and spectral correlation. To achieve this, the benchmark hyperspectral dataset with CNN deep learning model for hyperspectral image classification is taken and analyzed for different learning rates for SGD optimizer. The accuracy percentage were maximum when the SGD optimizer were trained with maximum learning rate. As the learning rate decreases, so does the accuracy percentage.

Agriculture and its allied activities play an important role in the economy of a nation. The ever growing population of the world and the unexpected changes occurring in environment makes it necessary to replace the age old agricultural practices with climate sustainable agriculture practices. The emergence of Artificial Intelligence (AI) and its allied fields have opened new avenues in the field of agriculture. Recent hybridization of IoT (Internet of Things) with AI fields has enabled the merging of intelligence in network of devices connected through internet facilitating availability and efficient analysis of huge amounts of real-time agriculture data. The present study is done with following objectives: (i) to study the role of AI and its related fields in agriculture, and (ii) to explore the future of various hybridized techniques being used in different areas of agriculture.