Sarojananda Mishra

An adaptive local linear optimized radial basis functional neural network model for financial time series prediction

Abstract For financial time series, the generation of error bars on the point of prediction is important in order to estimate the corresponding risk. In recent years, optimization techniques-driven artificial intelligence has been used to make time series approaches more systematic and improve forecasting performance. This paper presents a local linear radial basis functional neural network (LLRBFNN) model for classifying finance data from Yahoo Inc. The LLRBFNN model is learned by using the hybrid technique of backpropagation and recursive least square algorithm. The LLRBFNN model uses a local linear model in between the hidden layer and the output layer in contrast to the weights connected from hidden layer to output layer in typical neural network models. The obtained prediction result is compared with multilayer perceptron and radial basis functional neural network with the parameters being trained by gradient descent learning method. The proposed technique provides a lower mean squared error and thus can be considered as superior to other models. The technique is also tested on linear data, i.e., diabetic data, to confirm the validity of the result obtained from the experiment.

A self-adaptive fuzzy-based optimised functional link artificial neural network model for financial time series prediction

In recent years, new data mining and machine learning techniques have been developed and applied to various fields of science. Out of these recently developed techniques few offer online support and are able to adapt to large and complex financial dataset. Therefore, the present research adopts Functional Link Artificial Neural Network (FLANN) model for predicting the closing price of three companies namely Yahoo Inc, Nokia and Bank of America. The FLANN model used is trained by fuzzy after normalisation of the data and closing price is forecasted for one day and one week ahead. The prediction result is compared with the parameters of the FLANN model trained by Genetic Algorithm (GA) and Particle Swarm Optimisation (PSO). The proposed training method provides better accuracy and takes less time as compared to training the FLANN model using PSO or GA. The proposed approach has also been compared with a linear dataset for validation. The FLANN-fuzzy approach is seen to provide better results in predicting financial distress.

A harmony search-based artificial neural network for stock market prediction

For financial time series, the generation of error bars on the point prediction is important in order to estimate the corresponding risk. In recent years, artificial intelligence optimisation techniques have been used to make time series approaches more systematic and improve forecasting performance. The harmony search learning methodology, already successfully applied for training of multilayer perceptrons, is applied to Functional Link Artificial Neural Network (FLANN) in order to infer non-linear models for predicting a time series and the related volatility. The proposed method is implemented and the results are compared with FLANN model trained by back propagation and differential evolution. The proposed training method shows that FLANN-harmony search provides better forecasting/prediction as compared to training the FLANN model using back propagation or differential evolution.

Secure Hybrid Routing for MANET Resilient to Internal and External Attacks

An ad hoc network is the cooperative engagement of a collection of mobile nodes without the required intervention of any centralized access point or existing infrastructure. Although there is an increasing trend to adopt ad hoc networking for commercial uses, their main applications lie in military, tactical and other security-sensitive operations. In these and other applications of ad hoc networks, secure routing is an important issue. Quite a good number of protocols have been suggested in this area, but most of them are either proactive or reactive in approach. Studies reveal that, either a pure proactive or a pure reactive approach of routing performs well in a limited region of network setting. However, in diverse applications of ad hoc networks the performance of either class degrades dramatically. In this paper, we have presented a Secure Hybrid Routing Protocol (SHRP) for MANET, which aims at addressing the above limitations by combining the best properties of both proactive and reactive approaches. The protocol is based on the design of zone routing protocol (ZRP). The paper details the design of the protocol and analyses its robustness in the presence of multiple possible security attacks against ad hoc routing caused either by an internal compromised node or an external advisory.

An Approach to Detect Patterns (Sub-graphs) with Edge Weight in Graph Using Graph Mining Techniques

The task of detecting pattern or sub-graph in a large graph has applications in large areas such as biology, computer vision, computer-aided design, electronics, intelligence analysis, and social networks. So work on graph-based pattern detection has a wide range of research fields. Since the characteristics and application requirements of graph vary, graph-based detection is not the only problem, but it is a set of graph-related problems. This paper proposes a new approach for detection of sub-graph or pattern from a weighted graph with edge weight detection method using graph mining techniques. The edge detection method is proposed since most of the graphs are weighted one. Hence this paper proposes an algorithm named EdWePat for detection of patterns or sub-graphs with edge weight detection rather node value.

A survey on women’s education and their economic condition in Odisha

Education has been considered as panacea for all sorts of social issues. This paper indents to study the education status and economic condition of women in the district of Khordha, Odisha. With the help of a survey instrument, data were collected from 446 respondents of rural and urban areas of Khordha district. Trained data collectors were appointed to collect data and the sampling method was multi stage purposive sampling method. Descriptive statistics of the collected data reveals that education level of the women have been increased substantially both in urban and rural areas, most of the respondents have good housing condition and access to clean drinking water, they have their own house. However, the employment status of women has not increased. The most important reason for school drop- out has been poverty, parental decision and early marriage. Further, there is a need for subsequent study in the other districts with reference to similar parameters so that a complete understanding of the inter district position of Odisha as a whole can be made. However, it is quite clear that despite inadequate formal employment women have made significant contribution towards productivity, economic well-being as well as social harmony. This immense contribution has made our society inclusive, sensitive and empathetic.

A review on application of artificial neural network(ANN) on protein secondary structure prediction

Soft computing technologies are the most efficient technology in the field of Bioinformatics now a days. So many researchers are very keen interested in this topic to find out many unknown facts applying soft computing techniques in the field of DNA RNA alignment, Protein Structure Prediction, Gene mapping etc. Artificial neural network is one of the good techniques for the protein structure prediction problem. Protein structure prediction is the prediction of secondary, territory and quaternary structures of protein molecules in the effect of any external agents. There are many researchers working upon the field of applying ANN technology in Protein Structure prediction. This paper provides a review upon such type of researches that are carried out over the years i.e. application of ANN in the field of Protein Structure Prediction. In this review article we have discussed various technologies like MLP, RBF, SVM, RBFN, PNN and their applications that are being carried out in several research papers. In the summary section it provides a clean comparison of research work carried out by several researches.

Self Similarity Effect of RTT and RTO in Network Congestion Control Mechanism

Round Trip Time (RTT) behavior of TCP causes self similarity. In Network traffic we observe Multi fractal nature due to RTO. The mechanism of fast recovery and fast retransmits improved and reduces the performances in network traffic. It is difficult to suggest an appropriate RTO algorithm for TCP, which mostly affect data transfer delay and congestion control. In general TCP transmits number of packets in the ON period and OFF period is approximately equivalent to RTT behavior in the network data flow. In this paper we shows self similarity between short RTT and high RTT also focus on RTT and RTO behavior. We also observe Role of RTT behavior, when load is high in network which causes Heavy-tailed effect. We simulate by taking different scenario in each scenario we put some CPEs and routers and put the result in this paper. We carried out the simulation using NetSim6.2..

Localization Adaptive Secure Routing for Ad-Hoc Cloud Networks Based on Synchronized Timestamp Approach

Ad-hoc cloud computing now a days has become one of the most talked about technologies in terms of cost and computational promises it present to the enterprise. The foundation of ad-hoc clouds relies on the fact that the cloud setup distributes itself over resources harvested from mobile devices in operation. In such a scenario, routing between the devices is comparable to that of a mobile ad-hoc network and hence accounts for similar level of security risks. This paper proposes a secure routing protocol applicable to ad-hoc cloud environment. Our protocol is based on timestamp approach synchronized at devices level that forms the cloud and optimizes itself depending upon the localization factor. We specify the design of the proposed protocol and examines its capacity to withstand the possible security threats.