Objectives: This study aims about the development of Anyuak language named entity recognition of its first kind. NER is a fundamental sub task in natural language processing and the high accuracy competence in NER system marks the effectiveness of the downstream tasks. Anyuak language named entity recognition concern is addressed by using a long short-term memory model to categorize tokens into predefined classes. Methods: A long short-term memory is used to model the NER for Anyuak language to detect and classify words into five predefined classes: Person, Time, Organization, Location, and Others (non-named entity words). Because of feature selection plays a vital role in long short-term memory framework, the experiment in this work were conducted to discover most suitable features for Anyuak NER tagging task. Findings: When we evaluated the experiment in cross-validation, we achieved a promising result of precision, recall, and F1-measure values of 98%, 90, and 94% respectively. From the experimental result, it is possible to determine that tag context, word features, part of speech tags, suffixes and prefixes are significant features in named entity recognition and classification for Anyuak language. Novelty: Finally we have contributed a new architecture for Anyuak NER which uses automatically features for Anyuak named entity recognition which are not dependent on other NLP tasks. We proved that deep learning models can be extended, trained and can work for Anuak languages.
Objectives: Agriculture is the main food source and farmers are challenging a great production loss annually due to plant leaf disease. Early identification of tomato plant diseases help farmers to take preventive measure to reduce production loss. As a result, to recognize tomato plant leaf diseases in its early stage, a deep learning approach is discussed. Methods: For tomato disease identification and classification a convolutional neural network model is used in this study. CNN is capable for fine-grained disease identification as a technique which avoids feature engineering and threshold segmentation through automatic feature extraction. Findings: In this experiment, we have used 22,930 leaf image dataset are taken from plant village dataset, some are collected from Awash Melkasa tomato cultivation area in various seasons. Image processing is conducted along with pixel with operations it enhance the image data followed with feature extraction of patterns of collected leaves to detect the leaf diseases. The extracted patterns are fit into the neural network model with 100 epochs, 80/20 splitting ratio, and 0.001 learning rate. Hence the tomato disease network model achieves an overall 98.3% accuracy performance. Novelty: In order to detect tomato leave disease, we performed image processing with pixel-wise operation to enhance the leaf images that can be followed by feature extraction to classify patterns. We extend, and adopt neural network using local images collected under challenging environment datasets and optimization is performed in Adam optimizer with categorical entropy as loss function.
Background/Objectives: Agriculture is a major food source for Ethiopian population. Plant diseases contribute a great production loss, which can be addressed with continuous monitoring. Early plant disease identification using computer vision and Artificial Intelligence (AI) helps the farmers to take preventive course of action to increase production quality. Manual plant disease identification is strenuous and error-prone. Methods: In this study, we present a convolutional neural network architecture inception-v3 model to detect potato leaf diseases using a deep learning-based transfer learning technique. We used separable convolution in the inception block that can minimize the number of parameters by an outsized margin and to utilize resource efficiently. The inception-V3 model have a higher training accuracy and needs less training time than the main CNN architecture, as the used parameters are fewer. Findings: In this study, there is an improvement on the little noisy on sample images which leads to misidentification of diseases. In our experiment, we have used an RGB color channel image dataset to train model, which yields an overall accuracy performance of 98.7% on the heldout test set. Novelty: In order to identify potato leave diseases, we conducted transfer learning for high performance classification with pixel-wise operation to enhance the number of leaf images. A model based on inception-v3 transfer learning approach is presented in this study for disease identification of potato leave images, thus provide an effective computer-aided recognition model for potato disease classification in the absence of large data.
Background/Objectives: Information has become part of our existence and to access the information from database we need to be skilled with database query languages such as SQL. Hence in this study we propose Amharic Language Interface to Database (ALIDB). Here, the request is simple like asking a human to do so in a local language (Amharic). Method: So far, different techniques such as pattern matching, syntax based, semantic grammar based and Intermediate Representation Language systems have been used to develop NLIDB. Among these techniques, the study employed pattern matching and similarity checking for developing Amharic language text retrieval from the Database. Findings: The result of the experiment shows 91% accuracy. However, the scheme has no impact on Amharic temporal queries. Further development will be done on the algorithm that includes temporal queries in ALIDB. Novelty: Finally we identified 20 rules and thereby contributed a new pattern / algorithm for this language that converts Amharic sentence into a Structured Query Language (SQL) and fetch results from the Database.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.