Combination Of K-nearest Neighbor Research Articles

Inkjet printer prediction under complicated printing conditions based on microscopic image features

A novel technique is introduced to predict the printer model used to produce a given document. Samples containing only a few letters printed under varying conditions (i.e., different printing modes, letter types, fonts) were collected to establish a dataset of 41 inkjet printer models from common manufacturers, such as HP, Canon, and Epson. Morphological features were analyzed by extraction of image features using several algorithms in a series of microscopic images and a Wilcoxon test was used to measure the significance of variations between printed samples. Significant differences between various printing conditions might post potential challenge to questioned document examination. Discriminant analysis and the k-nearest neighbor (KNN) algorithm were also employed for source printer prediction under varying printing condition on 30% images with the rest images as training dataset. The results of a validation experiment demonstrated that while quadratic discriminant analysis (QDA) achieved an accuracy of 96.3%, a combination of KNN and QDA reached 98.6%. As such, this technique could aid in the forensic examination of printed documents.

Performance of Differential Evolution Algorithms for Indoor Area Positioning in Wireless Sensor Networks

In positioning systems in wireless sensor networks, the accuracy of localization is often affected by signal distortion or attenuation caused by environmental factors, especially in indoor environments. Although using a combination of K-Nearest Neighbor (KNN) algorithm and fingerprinting matching can reduce positioning errors due to poor signal quality, the improvement in accuracy by increasing the number of reference points and K values is not significant. This paper proposes a Differential Evolution-based KNN (DE-KNN) method to overcome the performance limitations of the KNN algorithm and enhance indoor area positioning accuracy in WSNs. The DE-KNN method aims to improve the accuracy and stability of indoor positioning in wireless sensor networks. According to the simulation results, in a simple indoor environment with four reference points, when the sensors are deployed in both fixed and random arrangements, the positioning accuracy was improved by 29.09% and 30.20%, respectively, compared to using the KNN algorithm alone. In a complex indoor environment with four reference points, the positioning accuracy was increased by 32.24% and 33.72%, respectively. When the number of reference points increased to five, in a simple environment, the accuracy improvement for both fixed and random deployment was 20.70% and 26.01%, respectively. In a complex environment, the accuracy improvement was 23.88% and 27.99% for fixed and random deployment, respectively.

An improved k-NN anomaly detection framework based on locality sensitive hashing for edge computing environment

Large deployment of wireless sensor networks in various fields bring great benefits. With the increasing volume of sensor data, traditional data collection and processing schemes gradually become unable to meet the requirements in actual scenarios. As data quality is vital to data mining and value extraction, this paper presents a distributed anomaly detection framework which combines cloud computing and edge computing. The framework consists of three major components: k-nearest neighbors, locality sensitive hashing, and cosine similarity. The traditional k-nearest neighbors algorithm is improved by locality sensitive hashing in terms of computation cost and processing time. An initial anomaly detection result is given by the combination of k-nearest neighbors and locality sensitive hashing. To further improve the accuracy of anomaly detection, a second test for anomaly is provided based on cosine similarity. Extensive experiments are conducted to evaluate the performance of our proposal. Six popular methods are used for comparison. Experimental results show that our model has advantages in the aspects of accuracy, delay, and energy consumption.

Crop Recommendation System using hybrid of KNN and Random Forest Classifier

Machine learning and its rapid advancement have significantly improved the way we interact with computers. We can find applications of machine learning in almost every field, like the IT industry, medicine, agriculture, etc. The idea of imparting machine learning to agriculture rose decades ago, and, as a result, many improvements were made in the field of agriculture. Various models are developed to predict the crop and yield using machine learning algorithms like decision trees, but the main problem with using algorithms like decision trees is that they do not provide the desired accuracy, which may lead to incorrect predictions. This paper proposes a user-friendly crop recommendation and yield prediction system. The user provides the following as input: state name, district name, soil type, and season. To recommend the crop and predict the yield of the crop, a combination of K-nearest neighbor (KNN) and random forest (RM) is used. The K-nearest neighbor algorithm showed 98% accuracy, and the Random Forest algorithm showed 96% accuracy

Klasifikasi Jenis Buah Jambu Biji Menggunakan Algoritma Principal Component Analysis dan K-Nearest Neighbor

The maturity level of guava fruit can be determined by looking at various factors. Shape is one of the factors that play a role in identifying certain objects. The classification of guava fruit can be seen from the shape, texture and color. The shape of the guava fruit is quite diverse ranging from round (Round shape) to oval (Pear shape). So a Matlab application was built to determine the type of guava based on its color, shape and texture. K-Nearest Neighbor can classify objects based on learning data that is closest to the object so that the results can be more accurate. Principal Component Analysis (PCA) is a statistical technique for simplifying many-dimensional data sets into lower dimensions (extration features). The combination of K-Nearest Neighbor with Principal Component Analysis produces a fairly high accuracy for determining the type of guava using a total of 45 images and divided into two data including training data with a total of 36 guava data and test data with a total of 9 guava data.

Palm Oil Maturity Classification Using K-Nearest Neighbors Based on RGB and L*a*b Color Extraction

This study aims to classify the maturity level of oil palm using the K-Nearest Neighbors (KNN) method based on the extraction of RGB and L*a*b color features. This classification determines the optimal oil production based on the color of the palm. However, this process is often carried out by humans, so sorting is not optimal because it is done manually, and it takes a long time. Thus, we propose automatic detection using machine learning based approach. In this study, the image processing methods are used to classify the level of oil palm maturity. The process starts with taking 150 photos with a smartphone. Each of them is then grouped into 3 classes, namely ripe (50), intermediate (50), and unripe (50). The oil palm image is pre-processed using the segmentation method (removing background), resizing, and cropping. Then, color features using RGB and L*a*b are extracted. Each of these extracted color features becomes input in the training and testing process of the KNN algorithm and is validated using k-fold cross-validation, with the results of accuracy being 95.4% and 97.4%. The combination of KNN and L*a*b color extraction approaches has better accuracy than KNN and RGB in terms of model accuracy (by 2–4.8 percentage points).

Recommendation System Using the K-Nearest Neighbor Approach: A Case Study of Dual Camera Quality as a Smartphone Selection Criterion

Many smartphones today need to be more precise about choosing one that suits the user's needs. In fact, smartphone sellers sometimes need help recommending smartphones that suit buyers' needs. Generally, buyers search for smartphone specifications with keywords they desire, but the results appear different from what they expected. Users need the main specifications, such as Random Access Memory (RAM) and Read Only Memory (ROM) capacity, battery, and high camera quality. This research aims to implement the K-Nearest Neighbor (KNN) algorithm for recommendation smartphone selection based on the criteria mentioned. The data test results show that the combination of KNN with four criteria has good performance, as indicated by the accuracy, precision, recall, and f-measure values of 95%, 94%, 97%, and 95%, respectively.

Classification of Broadleaf Weeds Using a Combination of K-Nearest Neighbor (KNN) and Principal Component Analysis (PCA)

Palm oil is one of the leading commodities in Indonesia. Oil palm yields can be influenced by several factors, one of which is proper weed control. Uncontrolled weeds can damage oil palm plantations. To be able to manage and control weeds, especially large leaf weeds, it is necessary to know the types of weeds. However, not all farmers have knowledge about the types of weeds. For that we need a system that can help identify broadleaf weeds based on leaf images using image processing. So this study aims to build a large leaf weed classification system using a combination of the K-Nearest Neighbor (KNN) and Principal Component Analysis (PCA) algorithms. PCA is used as feature extraction based on the characteristics formed from each spatial property. PCA can be used to reduce and retain most of the relevant information from the original features according to the optimal criteria. The results of the information will then be used by KNN for learning by paying attention to the closest distance from the object. Based on the test results, the developed model is able to produce an accuracy of 90%. Principal Component Analysis (PCA) and K-Nearest Neighbor (KNN) algorithms can be used in the classification process properly. Accuracy results are strongly influenced by the amount of training data and test data as well as the quality of the image used.

Assessment and prediction of acute subdural hematomas in vehicle collision accidents

Acute subdural hematomas (ASDHs) caused by vehicle collisions are common brain injuries that suffers prone to disability and death. However, there is currently no commonly accepted evaluation index for ASDHs, such as the head injury criterion (HIC) for evaluating skull injuries and the brain injury criteria (BrIC, BRIC) for evaluating diffuse axonal injuries. Therefore, we established a set of ASDH indexes (ASDHIs) based on the genetic algorithm for different biomechanical metrics and developed an ASDH prediction model based on an artificial neural network learning (ANN-L). First, simulation models were built using 218 sets of crash test data and the simulated injury monitor model (SIMon). We calculated the correlation coefficients between the kinematic parameters, such as the translational and rotational acceleration, and the biomechanical metrics, such as the intracranial pressure and strain of the bridging vein. The effect of some outlier data is removed using a combination of K-Nearest Neighbor and quadratic regression. We screened some parameters with strong correlations and calculated their weight factors based on the genetic algorithm to establish a set of evaluation indexes for ASDHs. Comparing with 19 existed injury criteria, the results show that each established ASDHI has a higher correlation than these traditional criteria, and also indicate that the evaluation performance of each established ASDHI is advantageous. Finally, on the basis of these established ASDHIs, we built an accurate prediction model of ASDHs using ANN-L, which makes it possible to realize the prediction of ASDHs by only depending on kinematic responses.

A Hybrid Model to Predict Monthly Streamflow Using Neighboring Rivers Annual Flows

The issue of predicting monthly streamflow data is an important issue in water resources engineering. In this paper, a hybrid model was proposed to generate monthly streamflow data for a river from the annual streamflow data. To develop the proposed hybrid model, a combination of K-Nearest Neighbor (KNN) and Random Tree (RT) algorithms was used. The hybrid structure was designed to predict the annual flow data for the target river using the annual flow data from neighboring rivers by applying the KNN model, and then generated the monthly flow data for the target river using the predicted annual flow by applying the random tree model. The hybrid model was applied to three rivers in Iraq. The accuracy of the proposed model was tested using two statistical indices, namely, the degree of determination and the efficiency coefficient. The results of the statistical indices indicated a good performance of the proposed hybrid model to generate monthly streamflow using annual streamflow data that the values of degree of determination and efficiency coefficient were greater than 0.91 in the training phase and greater than 0.79 in the validating phase.

Multipurification of matching pairs based on ORB feature and PCB alignment case study

To address the scale invariance and the mismatching problems of ORB (oriented FAST and rotated BRIEF), an improved algorithm based on multipurification was put forward and applied in PCB matching and positioning. ORB feature points were initially extracted from both the template image and the target image. Rough matching was then performed by the combination of k-nearest neighbor (k-NN) algorithm and best bin first search algorithm. Considering the existence of a large number of mismatches, the multipurification that consists of neighborhood ratio, bidirectional matching, and cosine similarity was used to purify the set of matching pairs. Finally, the affine matrix between the two images was solved by progressive sample consensus. The experimental results showed that the matching accuracy was significantly improved after the purification by the proposed method. The improved algorithm outperformed original ORB under the condition of image rotation and scaling. The mean rotation angle error was just 1/10th of that of original ORB. The time overhead of the improved algorithm was comparable to that of original ORB, which was about six times faster than that of speeded up robust features and 30 times faster than that of scale-invariant feature transform. The enhanced algorithm demonstrated great advantages in the improvement of accuracy, time overhead, and robustness for the PCB alignment.

Short-Term Traffic Flow Forecasting Based on Combination of K-Nearest Neighbor and Support Vector Regression

To improve the forecasting accuracy of short-term traffic flow and provide precise and reliable traffic information for traffic management units and travelers, this study proposes a hybrid prediction model that is based on the characteristics of K-nearest neighbor (KNN) method and support vector regression (SVR). The proposed hybrid model, i.e. KNN-SVR, mimics the search mechanism of the KNN method to reconstruct a time series of historical traffic flow that is similar to the current traffic flow. Then, the SVR is used for short-term traffic flow forecasting. Using actual traffic flow data, we study the effect of the traffic flows on target and adjacent section roads and analyze the forecasting accuracy of the proposed model. Results show that the KNN-SVR model that considers the target and adjacent section roads has the best performance, having a mean absolute percentage error (MAPE) of 8.29%. The forecasting error of the KNN-SVR model that considers only the target section road is slightly large, having an average MAPE of 9.16%. Furthermore, the forecasting accuracy of the KNN-SVR model is better than that of traditional prediction models, such as the KNN method, SVR, and neural networks.

Utilization of data classification in the realization of a surface Plasmon resonance readout system using an FPGA controlled RGB LED light source

This paper presents the realization of a surface plasmon resonance (SPR) sensor readout system using a tricolor red, green, and blue light emitting diode light source. Time domain intensity modulation of each color channel is applied to interrogate three bands of interest in the SPR spectrum using a single photodiode detector. A low computing resource classification approach is used through the combination of k-nearest neighbor (kNN) and adapted clustering using representative. An optimized number of representatives is chosen in the validation process to reduce the required amount of data for the kNN classification. This scheme was used to classify the concentrations of different glucose solutions. The sensor readout system hardware is based on the use of a field programmable gate array and the glucose solution classification is developed and undertaken on a personal computer using the Python open source programming language.

Development of a Structured Algorithm to Identify the Status of a Machine Tool to Improve Energy and Time Efficiencies

Energy efficiency is a matter of concern for manufacturing industries due to ever increasing energy costs and strict environmental policies. Efficient monitoring of energy consumption of machine tools is first step towards energy conservation. This study presents a non-intrusive energy monitoring technique to quantify the energy consumption of a machine tool at the unit process level and determines the operational status of the machine tool using the energy data profile. A combination of K-nearest neighbors and principal component analysis is used to develop smart energy sensor which can determine the time and energy spent during each operational state of the machine tools. It also identifies the operational state of various machine tool components such as spindle, coolant pump, automatic tool changer, etc. The time and energy map provided by the proposed sensor will help the practitioners to identify the potential areas of energy and time saving.

A COMBINATION DEEP BELIEF NETWORKS AND SHALLOW CLASSIFIER FOR SLEEP STAGE CLASSIFICATION

In this research, it is proposed to use Deep Belief Networks (DBN) in shallow classifier for the automatic sleep stage classification. The automatic classification is required to minimize Polysomnography examination time because it needs more than two days for analysis manually. Thus the automatic mechanism is required. The Shallow classifier used in this research includes Naïve Bayes (NB), Bayesian Networks (BN), Decision Tree (DT), Support Vector Machines (SVM), and K-Nearest Neighbor (KNN). The results obtained that many methods of the shallow classifier are increasing precision, recall, and F-Measure if they use DBN output as input for classification. Experiments that have been done indicate a significant increase of Naive Bayes after being combined with DBN. The high-level features generated by DBN are proven to be useful in helping Naive Bayes' performance. On the other hand, the combination of KNN with DBN shows a decrease because high-level features of DBN make it harder to find neighbors that optimize the performance of KNN.

K-Nearest Neighbor Neural Network Models for Very Short-Term Global Solar Irradiance Forecasting Based on Meteorological Data

This paper proposes a novel methodology for very short term forecasting of hourly global solar irradiance (GSI). The proposed methodology is based on meteorology data, especially for optimizing the operation of power generating electricity from photovoltaic (PV) energy. This methodology is a combination of k-nearest neighbor (k-NN) algorithm modelling and artificial neural network (ANN) model. The k-NN-ANN method is designed to forecast GSI for 60 min ahead based on meteorology data for the target PV station which position is surrounded by eight other adjacent PV stations. The novelty of this method is taking into account the meteorology data. A set of GSI measurement samples was available from the PV station in Taiwan which is used as test data. The first method implements k-NN as a preprocessing technique prior to ANN method. The error statistical indicators of k-NN-ANN model the mean absolute bias error (MABE) is 42 W/m2 and the root-mean-square error (RMSE) is 242 W/m2. The models forecasts are then compared to measured data and simulation results indicate that the k-NN-ANN-based model presented in this research can calculate hourly GSI with satisfactory accuracy.

HydDB: A web tool for hydrogenase classification and analysis.

H2 metabolism is proposed to be the most ancient and diverse mechanism of energy-conservation. The metalloenzymes mediating this metabolism, hydrogenases, are encoded by over 60 microbial phyla and are present in all major ecosystems. We developed a classification system and web tool, HydDB, for the structural and functional analysis of these enzymes. We show that hydrogenase function can be predicted by primary sequence alone using an expanded classification scheme (comprising 29 [NiFe], 8 [FeFe], and 1 [Fe] hydrogenase classes) that defines 11 new classes with distinct biological functions. Using this scheme, we built a web tool that rapidly and reliably classifies hydrogenase primary sequences using a combination of k-nearest neighbors’ algorithms and CDD referencing. Demonstrating its capacity, the tool reliably predicted hydrogenase content and function in 12 newly-sequenced bacteria, archaea, and eukaryotes. HydDB provides the capacity to browse the amino acid sequences of 3248 annotated hydrogenase catalytic subunits and also contains a detailed repository of physiological, biochemical, and structural information about the 38 hydrogenase classes defined here. The database and classifier are freely and publicly available at http://services.birc.au.dk/hyddb/

Short Term Prediction of Freeway Exiting Volume Based on SVM and KNN

ABSTRACT In order to better predict the traffic states on freeways and make management decisions, a hybrid model of support vector machine (SVM) and K-nearest neighbor (KNN) is proposed for short- term freeway exiting volume prediction. First, a historical data set is built by using the freeway toll data. The abnormal toll records, such as records that have same entry and exit station, illogical time record and abnormal travel speed, are excluded by data quality control. Based on the historical dataset, it is found that the exiting volume has periodical variation over time which provides the basis of the short-term prediction. Then, the historical data set is cross-classified into twelves groups based on the day of week and time of day. The prediction has been done for each group. Finally, the prediction is accomplished by the hybrid-model of SVM and KNN. The exiting volumes of previous time periods are used as the feature vector for KNN and SVM. Besides, a dynamic weight is adopted for the prediction of current time period based on the latest prediction accuracy of KNN and SVM. The model results indicate that the proposed algorithm is feasible and accurate. The Mean Absolute Percentage Error is under 10%. When comparing with the results of single KNN or SVM method, the results show that the combination of KNN and SVM can improve the reliability of the prediction significantly. The proposed method can be implemented in the on-line application of exiting volume prediction, which is able to consider different vehicle types.

Learning from Imbalanced Multi-label Data Sets by Using Ensemble Strategies

Multi-label classification is an extension of conventional classification in which a single instance can be associated with multiple labels. Problems of this type are ubiquitous in everyday life. Such as, a movie can be categorized as action, crime, and thriller. Most algorithms on multi-label classification learning are designed for balanced data and don’t work well on imbalanced data. On the other hand, in real applications, most datasets are imbalanced. Therefore, we focused to improve multi-label classification performance on imbalanced datasets. In this paper, a state-of-the-art multi-label classification algorithm, which called IBLR_ML, is employed. This algorithm is produced from combination of k-nearest neighbor and logistic regression algorithms. Logistic regression part of this algorithm is combined with two ensemble learning algorithms, Bagging and Boosting. My approach is called IB-ELR. In this paper, for the first time, the ensemble bagging method whit stable learning as the base learner and imbalanced data sets as the training data is examined. Finally, to evaluate the proposed methods; they are implemented in JAVA language. Experimental results show the effectiveness of proposed methods.
 Keywords: Multi-label classification, Imbalanced data set, Ensemble learning, Stable algorithm, Logistic regression, Bagging, Boosting

The Design of kNN Text Categorization on Storm Cluster

Try to handle a big data stream for information retrieval is just a beginning from TREC KBA2012. Storm is a free and open source distributed real-time computation system, which makes it easy to reliably process unbounded streams of data. For the issue of KBA2012, the combination of k-nearest neighbor (kNN) algorithm and Storm cluster will be an effective solution. kNN classification technique stands out for its simplicity and capability of being implemented on distributed platform. In addition, the entities (categories) in the issue of KBA2012 are settled down, which corresponds to the situation in which kNN is used. The paper discussed the implementation of kNN on Storm cluster and the results of experiments with KBA2012 data set.