About 15% of sugarcane leaf is defective because of diseases, it reduces the quantity and quality of sugarcane production significantly. Early detection and estimation of plant disease is a way to control these diseases and minimize the severe infection. This paper proposes a model to identify the severity of certain spot disease which appear on leaves based on segmented spot. The segmented spot is obtained by thresholding a* component of L*a*b* color space. Diseases spots are extracted with maximum standard deviation of segmented spot that use for detection the type of disease using classification techniques. The classifier is a Support Vector Machine (SVM) which uses L*a*b* color space for its color features and Gray Level Co-Occurrence Matrix (GLCM) as its texture features. This proposed model capable to determine the types of spot diseases with accuracy of 80% and 5.73 error severity estimation average.
Emotion detection is an application that is widely used in social media for industrial environment, health, and security problems. Twitter is ashort text messageknown as tweet.Based on content and purposes, the tweet can describes as information about a user"s emotion. Emotion detection by means oftweet, is a challenging problem because only a few features can be extracted. Getting features related to emotion is important at the first phase of extraction, so the appropriate features such as a hashtag, emoji, emoticon, and adjective terms are needed. We propose a new method for analyzing the linkages among features and reducedsemantically using Non-Negative M atrix Factorization (NM F). The dataset is taken from a Twitter application using Indonesian language with normalization of informal terms in advance. There are 764 tweets in corpus which have five emotions, i.e. happy (senang), angry (marah), fear (takut), sad (sedih), and surprise(terkejut). Then, the percentage of user"s emotion is computed by k-Nearest Neighbor(kNN) approach. Our proposed model achieves the problem of emotion detectionwhich is proved by the result near ground truth.
Segmentasi citra merupakan suatu metode penting dalam pengolahan citra digital yang bertujuan membagi citra menjadi beberapa region yang homogen berdasarkan kriteria kemiripan tertentu. Salah satu syarat utama yang harus dimiliki suatu metode segmentasi citra yaitu menghasilkan citra boundary yang optimal.Untuk memenuhi syarat tersebut suatu metode segmentasi membutuhkan suatu klasifikasi piksel citra yang dapat memisahkan piksel secara linier dan non-linear. Pada penelitian ini, penulis mengusulkan metode segmentasi citra menggunakan SVM dan entropi Arimoto berbasis ERSS sehingga tahan terhadap derau dan mempunyai kompleksitas yang rendah untuk menghasilkan citra boundary yang optimal. Pertama, ekstraksi ciri warna dengan local homogeneity dan ciri tekstur dengan menggunakan Gray Level Co-occurrence Matrix (GLCM) yang menghasilkan beberapa fitur. Kedua, pelabelan dengan Arimoto berbasis ERSS yang digunakan sebagai kelas dalam klasifikasi. Ketiga, hasil ekstraksi fitur dan training kemudian diklasifikasi berdasarkan label dengan SVM yang telah di-training. Dari percobaan yang dilakukan menunjukkan hasil segmentasi kurang optimal dengan akurasi 69 %. Reduksi fitur perlu dilakukan untuk menghasilkan citra yang tersegmentasi dengan baik. Kata kunci: segmentasi citra, support vector machine, ERSS Arimoto Entropy, ekstraksi ciri.
Twitter is a social media application, which can give a sign for identifying user emotion. Identification of user emotion can be utilized in commercial domain, health, politic, and security problems. The problem of emotion identification in twit is the unstructured short text messages which lead the difficulty to figure out main features. In this paper, we propose a new framework for identifying the tendency of user emotions using specific features, i.e. hashtag, emoji, emoticon, and adjective term. Preprocessing is applied in the first phase, and then user emotions are identified by means of classification method using kNN. The proposed method can achieve good results, near ground truth, with accuracy of 92%.
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.