In the current generation, a huge amount of textual documents are generated and there is an urgent need to organize them in a proper structure so that classification can be performed and categories can be properly defined. The key technology for gaining the insights into a text information and organizing that information is known as text classification. The classes are then classified by determining the text types of the content. Based on different machine learning algorithms used in the current paper, the system of text classification is divided into four sections namely text pre-treatment, text representation, implementation of the classifier and classification. In this paper, a BBC news text classification system is designed. In the classifier implementation section, the authors separately chose and compared logistic regression, random forest and K-nearest neighbour as our classification algorithms. Then, these classifiers were tested, analysed and compared with each other and finally got a conclusion. The experimental conclusion shows that BBC news text classification model gets satisfying results on the basis of algorithms tested on the data set. The authors decided to show the comparison based on five parameters namely precision, accuracy, F1-score, support and confusion matrix. The classifier which gets the highest among all these parameters is termed as the best machine learning algorithm for the BBC news data set.

A Comparative Analysis of Logistic Regression, Random Forest and KNN Models for the Text Classification

In the recent years, the categorization of text documents into predefined classifications has perceived a growing interest due to the growing of documents in digital form and needs to organize them. Text categorization is one of the extensively used for natural language processing (NLP) applications have achieved using machine learning algorithms. Text classification is a challenging researcher to find the best suitable structure and technique. Classification process done using manual and automatic classification. This research paper covers the preprocessing, feature extraction, different algorithms and techniques for text classification and finally evaluates the performance metrics for assessment.

A Review of State Art of Text Classification Algorithms

Gully erosion is a serious threat to the state of ecosystems all around the world. As a result, safeguarding the soil for our own benefit and from our own actions is a must for guaranteeing the long-term viability of a variety of ecosystem services. As a result, developing gully erosion susceptibility maps (GESM) is both suggested and necessary. In this study, we compared the effectiveness of three hybrid machine learning (ML) algorithms with the bivariate statistical index frequency ratio (FR), named random forest-frequency ratio (RF-FR), support vector machine-frequency ratio (SVM-FR), and naïve Bayes-frequency ratio (NB-FR), in mapping gully erosion in the GHISS watershed in the northern part of Morocco. The models were implemented based on the inventory mapping of a total number of 178 gully erosion points randomly divided into 2 groups (70% of points were used for training the models and 30% of points were used for the validation process), and 12 conditioning variables (i.e., elevation, slope, aspect, plane curvature, topographic moisture index (TWI), stream power index (SPI), precipitation, distance to road, distance to stream, drainage density, land use, and lithology). Using the equal interval reclassification method, the spatial distribution of gully erosion was categorized into five different classes, including very high, high, moderate, low, and very low. Our results showed that the very high susceptibility classes derived using RF-FR, SVM-FR, and NB-FR models covered 25.98%, 22.62%, and 27.10% of the total area, respectively. The area under the receiver (AUC) operating characteristic curve, precision, and accuracy were employed to evaluate the performance of these models. Based on the receiver operating characteristic (ROC), the results showed that the RF-FR achieved the best performance (AUC = 0.91), followed by SVM-FR (AUC = 0.87), and then NB-FR (AUC = 0.82), respectively. Our contribution, in line with the Sustainable Development Goals (SDGs), plays a crucial role for understanding and identifying the issue of “where and why” gully erosion occurs, and hence it can serve as a first pathway to reducing gully erosion in this particular area.

/pdf/hybrid-machine-learning-approach-for-gully-erosion-mapping-gkirpizf.pdf

Hybrid Machine Learning Approach for Gully Erosion Mapping Susceptibility at a Watershed Scale

In order to ensure the learning accuracy of machine learning algorithm and solve the problem of parameter adjustment in training model, a Xgboost parameter adjustment strategy based on grid search and k-fold cross validation is proposed, which improves the learning efficiency compared with the traditional grid search; based on the Bayesian network probability model, the Bayesian Optimization Xgboost model is built, which greatly reduces the time that is required for training the model and improves the performance on classification accuracy compared with traditional Xgboost model. In the simulation experiment, the training results of Xgboost model using two different parameter adjustment methods are compared. Xgboost based on Bayesian Optimization performs better than Xgboost using grid search and k-fold cross validation on both training accuracy and efficiency. By comparing the training results of different models, the optimal model is obtained. Feature importance of optimal model is analyzed.

Application and Improvement of Xgboost Algorithm Based on Multiple Parameter Optimization Strategy

Body weight and length are often used as a reference index in classifying the nutritional status of infants and toddlers. In Posyandu Anggrek in Limo, Depok, Indonesia, the anthropometric index calculation used is still manual, using the z-score list or WHO NCHS standard deviation found in KMS (Healthy Cards for Toddlers). Nutritional status is only seen based on the colors in the KMS without calculation or looking at the index in the anthropometric table. Naive Bayes, one of the methods used in classifying by calculating probability data, will be used in this study to measure the nutritional status of infants and children. The variables used in evaluating the nutritional status of infants and toddlers are data about body weight and body length. Based on the results of research and discussions conducted, 46 data about infants and toddlers are divided into 90% testing data and 10% training data by type of sampling using stratified sampling at RapidMiner. Based on the data obtained, the accuracy of Naive Bayes in classifying the nutritional status of infants and toddlers up to 75%.

Implementation of Naïve Bayes Classification Algorithm on Infant and Toddler Nutritional Status

Text classification is an essential advance in characteristic dialect processing. It very well may be performed utilizing different classification algorithms. Hadoop Map Reduce is widely utilized in text classification to perform classification on colossal measure of text data. However, Map Reduce required a ton of time to perform the tasks thereby increasing latency and since the data is distributed over the cluster it builds time and thus reducing processing speed. Also Hadoop utilizes long queue of code. Motivated by this, we propose a basic yet compelling machine learning method which uses Naive Bayes classifier for text data. In Machine Learning approach, the classifier is built automatically by learning the properties of categories from a set of pre-defined training data. Hence, it can process complex furthermore, multi assortment information in dynamic situations. Here we propose a naive bayes classifier which scales directly with number of indicators and data points which can be used for both binary and multiclass classification problems. We implemented the presented schemes using Machine Learning tool. The experimental results demonstrate the performance improvement in the classification technique.

Classification and Optimization Scheme for Text Data using Machine Learning Naïve Bayes Classifier

Online Social Networks (OSN) have progressed exponentially in terms of both participants and data being shared on the OSN. The OSN is more prevalent, and user profile on OSN has become the second identity in users’ life. The most popular and widely used OSN (i.e., Facebook, Google+, Twitter, Instagram, etc.) is built on a centralized architecture. The centralized OSN is designed around a centralized server, and users’ data are stored on the server. The centralized OSN poses the risk of a single point of failure, disclosure of sensitive information, and the user has little control over their sensitive data. The distributed OSNs are designed as a solution to single-point-of-failure and data privacy problems. However, it adduces inefficiencies in delivering essential functionalities such as unique identity distribution, data availability and access control services. This research work proposes a novel blockchain-based decentralized OSN (BDOSN) framework. The proposed BDOSN framework ensures data solitude and integrity of data. It also attains confirmable identity for the user and provides friends’ recommendations. Two smart contracts are designed to achieve useful functionalities in this research work. The potentiality of the proposed BDOSN scheme is implemented as an exemplification on solidity-Ethereum-blockchain using our dataset and publicly available data sets. The outcome of experiment illustrate the efficiency and validity of the proposed BDOSN.

BDOSN: Privacy-aware Blockchain based Decentralized OSNs

In Horticulture, an automatic fruit plucking systems may significantly cut labor expenses while improving picking efficiency. Currently, an autonomous fruit plucking system depends heavily on vision intelligence to identify fruit from the background. Most known approaches are being tested under steady environmental conditions, where the light illumination is relatively perpetual, and the fruits are unimpeded. However, the large-scale implementation of autonomous fruit picking systems is hampered because of the increased ramification of the surrounding environment, including fluctuations in illumination and pervasive occlusion. We propose a fruit recognition technique based on finetuned YOLOv5s to address fruit plucking challenges. The proposed technique detects Custard Apple, Pomegranate, and Waxberry fruits. First, fruit images from a real-world setting were gathered and preprocessed to create a dataset. The model is then trained on the dataset to precisely recognize the fruits. The experimental findings demonstrate that Precision, Recall, mAP 0.5, and mAP 0.5:0.95 were 90.4%, 86.4%, 89.4%, and 68.5%, respectively, outperforming the existing approaches and having significant implications for autonomous fruit picking systems in orchards.

A Fruit Detection Method for Vague Environment High-Density Fruit Orchards

Deep learning algorithms help achieve promising results in diagnosis of cancer patient. However, obtaining patient confidential data and medical images of cancer patient from medical hospitals and diagnosis centre is challenging thus preventing the deep learning usage in analysis of medical data of patient. Among woman, Breast cancer is prevailing cancer and it has become common health problem. The classification of Breast cancer focuses on identifying breast tumours sub-classes. Deep learning is among the current trends in computer science research and it enhances neural networks. The layers in neural networks are more, deep which allows higher abstraction levels and help to get accurate analysis of data. Google’s recently proposed architecture-EfficientNet, has less parameters which results in faster convergence for classification using ImageNet which showed that EfficientNet outperformed prevailing architectures such as DenseNet and ResNet. This research work include comparing the performance of EfficientNetV2-B0 architecture with DenseNet and ResNet architectures and few others for multiple class tumour classification on BreakHis dataset, which contains around 8000 histopathology images of breast that vary in magnification. This research work offers a methods that is based on transfer learning that improves existing architecture in multi-class classification by relying on pretrained DCNN trained on ImageNet dataset. Pretrained weights of EfficientNetV2-B0 are transferred as initial weights and the model is fine tuned in order to identify different benign and malignant tissue samples in tumor cells and classify them into multiclass. The model proposed has been tested with optimised hyperparameters irrespective of magnification of the images and was able to achieve accuracy upto 98%. The results surpass earlier studies in terms of accuracy across the performance metrics defined for CAD systems of breast cancer based on histological images.

Histopathological Image Classification of Breast Cancer using EfficientNet

Fruit features such as infection level are critical for farmers, manufacturers, and end users of fruit in a plant management system. The purpose of this experiment is to segment the fruits by distinguishing between healthy and infected regions and calculating the infection area. We mostly concentrate on Guava, Papaya, and Pomegranate because the infection seen on the outside is frequently present on the inside as well. Furthermore, it is widely consumed around the world. For our segmentation work, we employ a number of popular and current architectures and encoders. After comparing all of the models, we settled on three fruit models that scored 72.8 percent, 74.3 percent, and 80.4 percent on the dice on the Guava, Papaya, and Pomegranate datasets, respectively.

Venkatesh

Papers

Classification and Optimization Scheme for Text Data using Machine Learning Naïve Bayes Classifier

BDOSN: Privacy-aware Blockchain based Decentralized OSNs

A Fruit Detection Method for Vague Environment High-Density Fruit Orchards

Histopathological Image Classification of Breast Cancer using EfficientNet

Fruit Healthiness Detection and Filtering System