Weeds are unwanted plant or crops in the agriculture region which leads to primary pest problem in modern agriculture farming. In order to control area specific weed control on basis of ...classification and management of disease in farmland, hyperspectral images have been acquired from the satellite images in the remote sensing area. With different observation conditions and sensor characteristics, hyperspectral image classification based on spectral evolution simultaneously extracts the sets of spectral signatures of endmembers and maps the corresponding abundance maps from multiple spectral images. It then utilizes multiple supervised and unsupervised mechanisms for class-specific variations on weed and its diseases. Obviously mapping method degrades on accuracy of the coupling of the spectral evolution simultaneously. In this paper, a novel efficient weed classification and disease management on spectral evolution mapping should be proposed using Multivariate principle component analysis. It is examined as change detection mechanism which explores variation in the class features efficiently as the context of images is basis of bands of weed plant and its associated plant diseases, further it leads to a good tradeoff between wider receptive field and the use of Context is employed towards mapping Agriculture Land cover spectral evolution in the hyperspectral images. Proposed approach is capable of computing the spectral correlation among two images with respect to spectral similarity. Finally, it predicts the large intra class variation of weed accurately on temporal changes of the agriculture surfaces along various climate seasons and fields. Experimental analysis of the proposed mechanism was validated on Landsat 8 dataset to compute overall accuracy of the model on the changes in the weed and its diseases. The results of the work exhibits that proposed model can enhance the classification accuracy and reduces the differences of multi-temporal effects compared with existing state of art approaches.
The diagnosis of worn and damaged surfaces is an important issue in machine failure analysis and condition monitoring. Of many approaches used, image classification based on feature parameters has ...often proven to be particularly useful. However, large image databases can be computationally costly to analyse, and the datasets are susceptible to noise. Hence, it is essential to determine which feature parameters hold the most useful information, in order to improve the classification rate and computation time. This paper presents a performance evaluation of dimension reduction techniques currently used in pattern recognition. A comparison of three methods is conducted, in order to determine which is able to produce the best results over a large range of image datasets. The methods analysed are: Non-Linear Fisher, Principal Component Analysis and kernel Principal Component Analysis. These are then tested against four different classifiers to obtain the best combination. These classifiers are: Linear Discriminant Classifier, Quadratic Discriminant Classifier, k-Nearest Neighbour and Support Vector Machine Classifiers. For further analysis, two combined dimension reduction and classification methods are tested: Minimum Classification Error and reduced feature space Support Vectors. For the comparison, four datasets of images with different scales and rotations are used, i.e. Brodatz textures, artificially generated isotropic fractal images and Talysurf images of sandblasted and abraded steel surfaces. The results showed that a combination of the Non-Linear Fisher dimension reduction technique and a Linear Support Vector Machine Classifier gave the best performance overall and are the most promising for the application in automated machine condition monitoring and expert free failure analysis. Further improvement can be achieved by performing a step-wise dimension reduction by first reducing the features using the Principal Component Analysis method, then further reduction with the Non-Linear Fisher technique.
Random Forest based Intrusion Detection System for AMI M M, Savitha; Basarkod, P I
2022 IEEE Fourth International Conference on Advances in Electronics, Computers and Communications (ICAECC),
2022-Jan.-10
Conference Proceeding
The next generation in the electricity grid develops an Advanced Metering Infrastructure (AMI) with bi-directional communication. The Internet of Things (IoT) supports AMI to collect data from smart ...meters and identify the energy consumption pattern of consumers. A widely applied IoT routing protocol in AMI is Routing Protocol for Low-Power and Lossy Networks (RPL). The RPL protocol plays a vital role in ensuring a timely and reliable communication service on AMI. However, several factors affect the reliability of RPL in AMI, and those factors are security, delay, and unequal energy consumption between smart meters. To solve such issues, the Random Forest based Intrusion Detection System (IDS) (RFIDS) is designed for AMI. The proposed RF-IDS takes into account the results of offline traffic analysis using RF. However, it is insufficient to improve the RPL reliability since the design of RPL based only on offline traffic analysis does not necessarily mean that the designed protocol does not perform well online. Rather, the proposed IDS exploits an offline evaluation can be a first step to decide a security measure against several routing metrics. After offline traffic analysis, the RF-IDS can perform online traffic analysis to obtain delay and energy-related metrics. The offline traffic analysis process in RF-IDS selects an important feature set of the RPL protocol and improves the performance of RF over AMI. Moreover, the RF-IDS attempts to improve AMI communication reliability using a balanced energy factor and prolong the lifetime of smart meters in AMI. The efficiency of RF-IDS over AMI in terms of security, reliability, and energy is evaluated using the Cooja simulator.
The tremendous growth of internet supported by the extensive connectivity among systems within different networks has promoted the execution of class of unauthorized activities and has made their ...detection sophisticated. Preventing such unlawful acts are rarely possible and hence detecting them is an essential for ensuring the security of information systems. The paper presents a network intrusion detection via pair wise angular distance computation supported by genetic algorithm. The NSLKDD dataset is been used for training and testing of this supervised learning method. The information gain is used for attribute selection operation on NSLKDD dataset. The proposed methodology is expressing lower time and space complexity.
Giving credit to prospective debtor is determined by the existence of credit scoring. The accuracy of credit scoring to classify the debtor data is very important. The method that can be applied is ...classification and one of the classification method is decision tree. One of the decision tree algorithm that can be used is C4.5 algorithm. In this paper, the problem that discussed is how to increase the accuracy of C4.5 algorithm to predict credit receipts. The increasing accuracy is conducted by applying the Split Feature Reduction Model and Bagging Ensemble. Split Feature Reduction Model is applied in the preprocessing process which split datasets to the amount of n. In this paper, datasets split into 4 splits. Split 1 consists of 16 features, Split 2 consists of 12 features, Split 3 consists of 8 features, and Split 4 consists of 4 features. Then, C4.5 algorithm is applied to every splits. The best accuracy result by applying split feature reduction model with C4.5 algorithm is in Split 3 amount 73.1%. Then, the best accuracy results obtained by applying the split feature reduction model and bagging ensemble with C4.5 algorithm is in Split 3 amount 75.1%. In comparison to the accuracy of C4.5 algorithm stand alone, the applying of split feature reduction model and bagging ensemble obtained increased accuracy by 4.6%.
Music genres are helpful means for recommending songs of preferences by containing characteristics related to instruments, musical rhythms and harmonic structure and melodies of the song. Song ...listeners often face difficulty in finding desired tracks due to the vast volume of available music data. So, in this context, machine learning-based approaches can contribute in developing sophisticated method that can classify music genres and eventually building recommendation systems for online streaming services. In this paper, we propose an integrated framework that considers musical features from both time and frequency domain and after necessary preprocessing stages, incorporates into a boosting model for classification. We incorporate CatBoost as an ensemble learning model due to the obvious benefits of increased speed, reduced overfitting and the ability to assign greater weights to certain samples, and minimal variance sampling. We evaluated our proposed framework on a Bangla music dataset and discovered some noteworthy results that support the effectiveness of our proposed integrated model. A remarkable characteristic of such an integrated machine learning model, which is a significant contribution to the Bangla Music Industry in the era of Industrial Revolution 4.0, is its ability to analyze information from multidimensional data in a self-optimized approach with strong decision-making abilities.
Extraction from large no. of features (genes signatures) are the major issues in the prediction of cancer and its specific type identification using microarray datasets. Even though most classifiers ...predict the class (normal or cancerous) for various cancers, the accuracy of prediction still suffers. This is due to the importance of fewer gene signatures for a particular cancer and classification of samples independent from their originating form. The present paper proposes gene extraction techniques to work in an unsupervised manner. The proposed technique takes the advantage of both linear and non-linear feature extraction methods. Principal component analysis (PCA) is used in a linear manner whereas Denoising Autoencoder (DAE) is used in a nonlinear manner. In the first phase of the work, feature space extracts from both the methods have been combined and new features space has been utilized for cancer classification. Here, Four classifiers: Support vector machine (SVM), Multilayer perceptron (MLP), Naive Bays (NB) and Decision Tree are applied on a no. of gene signatures extracted from four different cancer datasets. It is seen that after feature extraction from this PCA-DAE, classification accuracy either increases or attains its maximum value in comparison with base techniques except NB.
During the course of the residency, novice surgeons develop specific skills before they perform actual surgical procedures. Manual feedback and assessment in basic robotic-assisted minimally invasive ...surgery (RMIS) training take up much of the expert surgeons' time, while it is very favorable to automatically feedback to all surgeons in various skill levels. Towards this end, we use the surgical robot kinematic dataset named JIGSAWS, a public database collected from Da Vinci robot operated by 7 surgeons, to extract 49 metrics for the suturing task using three types of features, namely time and motion-based, entropy-based, and frequency-based. To find out the most relevant metrics in skill assessment, we perform and compare two feature selection/reduction methods, namely principal component analysis (PCA) and relief algorithm. We separately reduce the features based on these two methods, while using a combination method. Although resulting in an acceptable accuracy of 84% when using each separately, the combination method results in 92% particularly noteworthy accuracy.
Feature reduction methods have been successfully applied to text categorization. In this paper, we perform a comparative study on three feature reduction methods for text categorization, including ...Document Frequency (DF), Term Frequency Inverse Document Frequency (TFIDF) and Latent Semantic Analyses (LSA). Our feature set is relatively large (since there are thousands of different terms in different texts files). We propose the use of the previous feature reduction methods as a preprocessor of Back-Propagation Neural Network (BPNN) to reduce the input data on training process. The experimental results on an Arabic data set demonstrate that among the three dimensionality reduction techniques proposed, TFIDF was found to be the most effective in reducing the dimensionality of the feature space.
In this paper, a novel global non-destructive evaluation (NDE) technique based on information fusion is proposed to diagnose loosening fault of clamping support. Two feature extraction methods are ...used to extract feature, which are wavelet package transform and power spectrum density analysis. During diagnosing loosening fault, two local decisions are made by using WP feature and PSD feature respectively. Then the two features are fused to make another local decision. Lastly, the three local decisions are fused to make global decision. The information fusion result have high correct diagnosis ratio and good antinoise performance. The correct diagnosis ratios with no noise and random noise reach 94.3% and 88.6% respectively.