Less Number Of Nodes Research Articles

A Novel Weights of Weights Approach for Efficient Transfer Learning in Artificial Neural Networks

Deep learning is often reputed to be slow at training due to large number of layers. There are several approaches including used for improving training mechanism. Transfer learning sometimes referred as transfer of knowledge is one such approach. However, Transfer learning is often moving weights or layers as a whole including weights that might not be significant.This research tries to implement a novel Weights of Weights approach where significant weights are extracted from a training network. Only the significant weights are transferred to new untrained network instead of transferring all the layers or weights. Experiments are carried out using a 7-layered neural networks on multivariate wine dataset to evaluate the efficiency of the model.The experiment results show that the transferable model has provided better accuracy of over 40%-60% (based on layers) without retraining compared to traditional approach of copying all the layers with same and less number of nodes.

Low energy consumption in manet network

The aim of this paper is design and develop energy efficient MANET network in wireless networks. One of the most significant and effective protocol based on low energy consumption and number of Ad-hoc is MANET as remote directing convention source nodes forward in network simulator. Less number of nodes in the network would give low energy usage or consumption as the nodes in the network exceeds or increases that will also increase the energy consumption in the network. The designed MANET system is tried with 9, 12, 15 and 18 number of nodes in a system using network simulation-2 (NS-2). Henceforth source node needs to restart over and over which brings about low energy consumption use and use, ectiveness is less and packet space is additionally less and throughput is likewise less and more start to finish delay. Arrangement of this issue in MANET convention which is advanced as the node doesn't advance when demand arrived at their first it checked there is low energy consumption (battery lifetime) and until the node energy consumption is more noteworthy than the limit. Designed MANET examinations of the energy consumption and node energy consumption by maintaining a strategic distance from the low number of nodes in a network. By contrasting energy consumption and node it demonstrates that MANET is far superior to existing framework 802.11 protocol convention based on battery lifetime, energy consumption, throughput, and power transmission. We have performed a comparison between EEM and AODV routing protocol considering different measuring parameters.

OAFPM: optimized ANFIS using frequent pattern mining for activity recognition

Fall causes serious perils for the elder people when they are living alone. A mathematical model, optimized ANFIS using frequent pattern mining (OAFPM) for activity recognition, is proposed in this paper, which uses fuzzy inference system, adaptive neural network and frequent pattern mining (FPM) to identify the activity of a person accurately. Accelerometer values are given as input to the proposed model in real time which forms the premise part of the model, whereas the consequence is defined by the rules generated out of input and output linear relation. Initial rule identification is done through membership functions of each activity, and the number of rules is reduced using FPM approach. During the learning phase, the optimal premise parameters are selected using gradient descent method and the choice of consequent parameters is based on the least-square estimation method. The optimal values of premise and consequent parameters along with the reduced rule matrix made the OAFPM model to achieve an accuracy rate of 95.8 $$\%$$ and also reduce the computational complexity by triggering less number of nodes for each activity.

Advancing Deep Learning to Improve Upstream Petroleum Monitoring

Data analytics is rapidly growing field in both academia and industry dealing with processing and interpreting large and complex data sets. It has got already many successful applications via advancing machine (ML) and deep learning (DL) techniques, starting to evolve in the upstream petroleum industry as well. The industry operates now with huge amount of sensors installed in different facilities, particularly in production and injection wells. These sensors provide millions of measurements, such as pressure, temperature, and rate every year for every well. The measurements may be highly correlated and carry crucial information for decision making. This paper concentrates on pressure-rate data sets accumulated with massive installation of permanent downhole gauges in such wells. The non-linear autoregessive (NARX) and the long short term memory (LSTM) neural networks were assembled and tested on a synthetic data set to compare results of pressure prediction, already addressed in the literature. The LSTM provided better predictions, but did not manage to capture entirely the pattern of the data. The shifting window method was then applied to improve the LSTM prediction capabilities, based on previous successful application in forecasting electricity demand. The method implies smooth transition from training to prediction improving network performance. The LSTM with the shifting window provided more accurate results for pressure prediction, and it was then successfully applied for rate prediction. Testing of different configurations of the LSTM network has shown that the pressure prediction performs well with less number of nodes in the hidden layers if compared with the rate predictions. Significant error decrease is achieved relatively fast (after 20 iterations) for both prediction tasks, making such predictions feasible for large data sets. The results provide basis for filling gaps in well monitoring data and short-term performance forecast, crucial tasks for decision making in all the industries operating with wells.

Survival and treatment outcomes of metaplastic breast carcinoma: Single tertiary care center experience in Pakistan.

Metaplastic breast carcinoma (MBC) is a rare disease with incidence of less than 1%. MBC present with a larger tumor size, less number of nodes involved, mostly undifferentiated triple negative tumors. We aimed to determine progression-free and overall survival and reported hospital-based incidence of MBC. A retrospective closed Cohort study elicited data of 42 patients with MBC from January 2008 to December 2013; followed till August 2016. Kaplan-Meier method was applied to compute overall and progression-free survival analysis. Cox Proportional hazard ratios were computed to assess associations between survival and independent variables. Hospital-based incidence of MBC was 1.92% (42/2187), 95% CI [1.41-2.56]. The median age at tumor diagnosis was 54 years (range, 25-81 years). Thirty-nine (92.9%) patients had Grade III tumor. The most common histopathology was squamous (69%). The median tumor size was 4.5 cm (range, 0.8-17 cm). Nineteen (45.2%) patients had nodal involvement at diagnosis. Four patients (9.5%) had metastatic disease at presentation. Hormone receptors were positive in 19 (45.2%) patients. Her-2 neu receptor was positive in 9 (19%) patients. Sixteen (38.1%) patients had triple negative disease. Neoadjuvant and adjuvant chemotherapy was received by 10 (31.25%) and 19 (45.2%) patients respectively. Both median progression-free and overall survival was 38 months. Five-year progression-free and overall survival was 79.5% and 76.3%, respectively. We report better survival outcomes when compared to series described earlier despite our patient population presenting mostly with high grade, large tumors, and half of them exhibiting nodal and hormonal involvement.

Adaptive localization and mapping with application to planetary rovers

AbstractFuture exploration rovers will be equipped with substantial onboard autonomy. SLAM is a fundamental part and has a close connection with robot perception, planning, and control. The community has made great progress in the past decade by enabling real‐world solutions and is addressing important challenges in high‐level scalability, resources awareness, and domain adaptation. A novel adaptive SLAM system is proposed to accomplish rover navigation and computational demands. It starts from a three‐dimensional odometry dead reckoning solution and builds up to a full graph optimization that takes into account rover traction performance. A complete kinematics of the rover locomotion system improves the wheel odometry solution. In addition, an odometry error model is inferred using Gaussian processes (GPs) to predict nonsystematic errors induced by poor traction of the rover with the terrain. The nonparametric GP regression serves to adapt the localization and mapping to the current navigation demands (domain adaptation). The method brings scalability and adaptiveness to modern SLAM. Therefore, an adaptive strategy develops to adjust the image frame rate (active perception) and to influence the optimization backend by including high informative keyframes in the graph (adaptive information gain). The work is experimentally verified on a representative planetary rover under a realistic field test scenario. The results show a modern SLAM systems that adapt to the predicted error. The system maintains accuracy with less number of nodes taking the most benefit of both wheel and visual methods in a consistent graph‐based smoothing approach.

An extended element free Galerkin method for fracture analysis of functionally graded materials

ABSTRACTAn extended element free Galerkin method (XEFGM) has been adopted for fracture analysis of functionally graded materials (FGMs). Orthotropic enrichments functions are used along with the sub-triangle technique for enhancing the Gauss quadrature accuracy near the crack, and the incompatible interaction integral method is employed to calculate the stress intensity factors. Numerical simulations have proved that XEFGM provides more accurate results by less number of nodes (DOFs) in comparison with the unenriched EFGM and other conventional methods for several FGM problems with different crack locations and loadings. The results have been compared with the reference results, showing the reliability, stability, and efficiency of present XEFGM.Received 9 June 2014 Accepted 17 September 2014.

Semi-asynchronous and distributed weighted connected dominating set algorithms for wireless sensor networks

Energy-efficient backbone construction is one of the most important objective in a wireless sensor network (WSN) and to construct a more robust backbone, weighted connected dominating sets can be used where the energy of the nodes are directly related to their weights. In this study, we propose algorithms for this purpose and classify our algorithms as weighted dominating set algorithms and weighted Steiner tree algorithms where these algorithms are used together to construct a weighted connected dominating set (WCDS). We provide fully distributed algorithms with semi-asynchronous versions. We show the design of the algorithms, analyze their proof of correctness, time, message and space complexities and provide the simulation results in ns2 environment. We show that the approximation ratio of our algorithms is 3ln(S) where S is the total weight of optimum solution. To the best of our knowledge, our algorithms are the first fully distributed and semi-asynchronous WCDS algorithms with 3ln(S) approximation ratio. We compare our proposed algorithms with the related work and show that our algorithms select backbone with lower cost and less number of nodes.

New Derivative Based Open Newton-Cotes Quadrature Rules

Some new families of open Newton-Cotes rules which involve the combinations of function values and the evaluation of derivative at uniformly spaced points of the interval are presented. The order of accuracy of these numerical formulas is higher than that of the classical open Newton-Cotes formulas. An extensive comparison of the computational cost, order of accuracy, error terms, coefficients of the error terms, observed order of accuracy, CPU usage time, and results obtained from these formulas is given. The comparisons show that we have been able to define some new open Newton-Cotes rules which are superior to classical open rules for less number of nodes and less computational cost with increased order of accuracy.

A Sweep Coverage Scheme Based on Vehicle Routing Problem

As an emerging coverage problem in wireless sensor networks, sweep coverage which introducing mobile sensors to cover points of interest within certain time interval can satisfy monitoring request in some particular application scenarios with less number of nodes than the conventional static coverage approach. In this work, aiming to support dynamical POI coverage and data delivery simultaneously, a novel sweep coverage scheme, named VRPSC(Vehicle Routing Problem based Sweep Coverage), is proposed by modeling the minimum number of required sensors problem in sweep coverage as a Vehicle Routing Problem (VRP). In VRPSC, an insertion algorithm is first introduced to create the initial scanning routes for POIs, and then the Simulated Annealing is employed to optimize these routes. The simulation results show that the VRPSC scheme achieves better performance than existing schemes. DOI: http://dx.doi.org/10.11591/telkomnika.v11i4.2380

An Efficient Clustering Technique for Deterministically Deployed Wireless Sensor Network

The deployment mechanism in wireless sensor networks (WSN) affects the coverage, connectivity, bandwidth, packet loss, lifetime of network, etc. features [1]. Depending upon the application, the sensor nodes are deployed in either random or deterministic fashion, accordingly WSN has different requirements and features. In this work, a new clustering technique, Energy Efficient Deterministic ClusterHead Selection Algorithm (E 2 DCH), is proposed for deterministically deployed WSN. Better coverage with less number of nodes, minimum traffic from nodes to base station, balanced energy consumption are the main features of E 2 DCH to improve life time of WSN. The proposed algorithm uses dynamic routing from nodes to respective cluster head by considering the number of nodes and residual node energy of all the involved nodes. It includes an efficient technique for reorganizing the clusters. Analysis and simulation results demonstrate the correctness and effectiveness of the proposed algorithm.

Energy Constrained Reliable Routing Optimized Cluster Head Protocol for Multihop under Water Acoustic Sensor Networks

Underwater acoustic sensor network is an emerging technique consisting of sensor nodes, and AUVs all working together to sense various phenomenon, converts the sensed information into digital data, store the digital data and communicate to the base stations through the intermediate nodes. Also Underwater Acoustic Sensor Networks are playing a main role in ocean applications. Unfortunately the efficiency of underwater Acoustic Sensor Networks is inferior to that of terrestrial sensor networks due to the long propagation delay, narrow bandwidth and high error rates. Also battery life and storage capacity of node is limited. Many routing protocols are proposed to improve the efficiency of Under Water Acoustic Sensor Networks. However their improvement is not enough, so there is a need of suitable routing protocol that consider all these limitations and makes communication in underwater network viable. In this paper, we propose a protocol called Reliable Routing Optimized Cluster Head (RROCH) protocol, a network coding approach for multihop topologies. We used performance metrics like packet delivery ratio, energy consumption, end-to-end delay and throughput of sensor nodes. LEACH, HMR-LEACH, LEACH-M are compared for their performance at different traffic conditions, number of nodes and depth. By analyzing our simulation results we found that RROCH protocol may be used for denser network with low traffic and HMR- LEACH protocol is suitable for higher traffic with less number of nodes.

Are we achieving the benchmark of retrieving 12 lymph nodes in colorectal carcinoma specimens? Experience from a tertiary referral center in India and review of literature

The number of lymph nodes (LNs) retrieved from a specimen of colorectal carcinoma may vary. Factors that can possibly affect LN yield are age of the patient, obesity, location of the tumor, neoadjuvant therapy, surgical technique and pathologist's handling of the specimen. The aim of our study is to look at lymph node retrieval from colorectal cancer (CRC) specimens in our hands and review the literature. From May 2010 to January 2011, a total of 170 colorectal carcinoma cases were operated in our institute. Type of the surgeries, lymph node yield was looked at. There were 103 (60.6%) males and 67 (39.4%) females. The commonest age group was 50-59 years (30.6%). The surgeries included 107 surgeries for rectal carcinoma (63%) and 63 surgeries for colonic carcinoma (37%). Sixty six (38.8%) cases had received preoperative chemoradiotherapy, whereas 104 (61.2%) cases were without adjuvant therapy. The total lymph node positivity (metastatic disease) was 44.7% .The overall mean lymph node yield was 12.68 (range 0-63; median 11). The mean lymph node harvest in the age group < 39 was 15.76 whereas, the lymph node harvest in the group more than 39 years old was 11.90. ( statistically significant; P=0.03). The mean lymph node yield from specimens of rectal cancers (10.30) was lower than the mean lymph node yield from specimens for colonic cancers (16.71);( statistically significant, P<0.01). There was also statistically significant difference between the mean LN yield in chemoradionaiive cases (14.63) and in the cases where neoadjuvant therapy was received, (9.59); P<0.01. Pathologist while assessing a specimen of CRC should aim to retrieve a minimum of 12 LN. Surgical expertise and diligence of the pathologists remain two main alterable factors that can improve this yield. Neoadjuvant or preoperative radiotherapy can yield in less number of nodes.

A new MLPG method for elastostatic problems

This paper introduces a novel meshless local Petrov–Galerkin (MLPG) method by presenting a new test function as a schema to solve the elasto-static problems. It is seen that the four ordinary MLPG methods can also be approached using the present test function. Both the moving least square (MLS) and the direct method have been applied to the method to estimate the shape function and to impose the essential boundary conditions. The results of three studied elasto-static cases; “two dimensional cantilever beam”, “first mode fracture of a center-cracked strip” and “edge-cracked functionally graded strip” show that by using less number of nodes, the present work gives sufficiently more accurate results. Meanwhile the method can also unify various kinds of MPLGs and one may conclude that the model is a good replacement for other common approaches.

Connectivity-Guaranteed Hybrid Topology Management Scheme for Improving the Operational Lifetime of 3-Dimensional Wireless Sensor Networks

The battery of limited energy operates wireless sensor nodes, it is important to increase the lifetime of the wireless sensor network. The aim of our work is to design a hybrid topology management scheme for improving the lifetime of the wireless sensor networks with guaranteed connectivity. The sensing range of the sensor node is smaller than the radio range and hence there is a large number of nodes within the coverage area of a node. Hence the network connectivity can be maintained with less number of nodes. By using this fact, we can save energy by switching off the transceiver, which consumes most of the energy. On the other hand, the sensor node spends most of the time in monitoring state and the transceiver is idle. In the idle state the transceiver consumes almost the same energy as in receiving state. Hence putting the transceiver in the sleep mode when it is idle can save the significant energy. We have designed a hybrid topology management scheme for 3-Dimensional sensor networks by making use of these facts. Our hybrid scheme improves the lifetime of the 3-Dimensional sensor network by a factor of 19.

In search of an optimization technique for Artificial Neural Network to classify abnormal heart sounds

Artificial Neural Network (ANN) finds use in classification of heart sounds for its discriminative training ability and easy implementation. The selection of number of nodes for an ANN remains an important issue as an overparameterized ANN gets trained along with the redundant information that results in poor validation. Also a larger network means more computational expense, resulting more hardware and time related cost. Therefore, a compact and optimum design of neural network is needed towards real-time detection of pathological patterns, if any from heart sound signals. This work attempts to (i) design a compact form of output layer with less number of nodes than output classes, (ii) select a set of input features that are effective for identification of heart sound signals using Singular Value Decomposition (SVD), QR factorization with column pivoting (QRcp) and Fisher's F-ratio, (iii) make certain optimum selection of nodes in the hidden layer for a more effective ANN structure using SVD and (iv) select and prune weights based on the concept of local relative sensitivity index (LRSI) for empirically chosen overparameterized ANN structure for phonocardiogram (PCG) classification. It is observed that the proposed techniques perform better in terms of reduction of model residues and time complexity for classifying 12 different pathological cases and normal heart sound.

Two-phase clustering process for outliers detection

In this paper, a two-phase clustering algorithm for outliers detection is proposed. We first modify the traditional k-means algorithm in Phase 1 by using a heuristic “if one new input pattern is far enough away from all clusters' centers, then assign it as a new cluster center”. It results that the data points in the same cluster may be most likely all outliers or all non-outliers. And then we construct a minimum spanning tree (MST) in Phase 2 and remove the longest edge. The small clusters, the tree with less number of nodes, are selected and regarded as outlier. The experimental results show that our process works well.