<span lang="EN-US">Because plant disease is main cause of most plants’ damage, improving prediction plans for early detection of plant where it has disease or not is an essential interest of decision makers in the agricultural sector for providing proper plant care at appropriate time. Clustering and classification algorithms have proven effective in early detection of plant disease. Making clusters of plants with similar features is an excellent strategy for analyzing features and providing an overview of care quality provided to similar plants. Thus, in this article, we present an artificial intelligence (AI) model based on k-nearest neighbors (k-NN) classifier and k-efficient clustering that integrates k-means with k-medoids to take advantage of both k-means and k-medoids to improve plant disease prediction strategies. Objectives of this article are to determine performance of k-mean, k-medoids and k-efficient also we compare k-NN before clustering and with clustering in prediction of soybean disease for selecting best one for plant disease forecasting. These objectives enable us to analysis data of plant that help to understand nature of plant. Results indicate that k-NN with k-efficient is more efficient than other in terms of inter-class, intra-class, normal mutual information (NMI), accuracy, precision, recall, F-measure, and running time.</span>

The concept of interfacing brains with robots/machines has been capturing human interests for a long time. The technology of the Brain-computer interface (BCI) has been aimed at building an interface between the brain and any electronic/electrical device (such as, smart home appliance, a wheelchair, and robotics devices) with the use of the electroencephalogram (EEG) that can be defined as a non-invasive approach for the measurement of the electrical potentials from the electrodes that have been placed on the scalp, produced by the activity of the brain. Over the past years, pattern classification was a highly challenging research field. Presently, the tasks of the pattern classification. In this paper, we chose motor imagery with the use of the single trial EEG signal, the SOM has been utilized to classify the signal processing algorithm ( FICA). In comparison to other algorithms of the EEG signal analyses. It has achieved a classification accuracy of up to 88.% in comparison with the other method where the reported accuracy has been 65%. The SOM classification algorithm has been fast, simple, efficient, and easy to use. It achieved satisfactory results at the BCI.