Baikunth Nath

A fusion model of HMM, ANN and GA for stock market forecasting

In this paper we propose and implement a fusion model by combining the Hidden Markov Model (HMM), Artificial Neural Networks (ANN) and Genetic Algorithms (GA) to forecast financial market behaviour. The developed tool can be used for in depth analysis of the stock market. Using ANN, the daily stock prices are transformed to independent sets of values that become input to HMM. We draw on GA to optimize the initial parameters of HMM. The trained HMM is used to identify and locate similar patterns in the historical data. The price differences between the matched days and the respective next day are calculated. Finally, a weighted average of the price differences of similar patterns is obtained to prepare a forecast for the required next day. Forecasts are obtained for a number of securities in the IT sector and are compared with a conventional forecast method.

Automatic Detection of Vascular Bifurcations and Crossovers from Color Retinal Fundus Images

Identifying the vascular bifurcations and crossovers in the retinal image is helpful for predicting many cardiovascular diseases and can be used as biometric features and for image registration. In this paper, we propose an efficient method to detect vascular bifurcations and crossovers based on the vessel geometrical features. We segment the blood vessels from the color retinal RGB image, and apply the morphological thinning operation to find the vessel centerline. Applying a filter on this centreline image we detect the potential bifurcation and crossover points. The geometrical and topological properties of the blood vessels passing through these points are utilized to identify these points as the vessel bifurcations and crossovers. We evaluate our method against manually measured bifurcation and crossover points by an expert, and achieved the detection accuracy of 95.82%.

Blood Vessel Segmentation from Color Retinal Images using Unsupervised Texture Classification

Automated blood vessel segmentation is an important issue for assessing retinal abnormalities and diagnoses of many diseases. The segmentation of vessels is complicated by huge variations in local contrast, particularly in case of the minor vessels. In this paper, we propose a new method of texture based vessel segmentation to overcome this problem. We use Gaussian and L*a*b* perceptually uniform color spaces with original RGB for texture feature extraction on retinal images. A bank of Gabor energy filters are used to analyze the texture features from which a feature vector is constructed for each pixel. The fuzzy C-means (FCM) clustering algorithm is used to classify the feature vectors into vessel or non-vessel based on the texture properties. From the FCM clustering output we attain the final output segmented image after a post processing step. We compare our method with hand-labeled ground truth segmentation of five images and achieve 84.37% sensitivity and 99.61% specificity.

Optimal design of neural nets using hybrid algorithms

Selection of the topology of a network and correct parameters for the learning algorithm is a tedious task for designing an optimal Artificial Neural Network (ANN), which is smaller, faster and with a better generalization performance. Genetic algorithm (GA) is an adaptive search technique based on the principles and mechanisms of natural selection and survival of the fittest from natural evolution. Simulated annealing (SA) is a global optimization algorithm that can process cost functions possessing quite arbitrary degrees of nonlinearities, discontinuities and stochasticity but statistically assuring a optimal solution. In this paper we explain how a hybrid algorithm integrating the desirable aspects of GA and SA can be applied for the optimal design of an ANN. This paper is more concerned with the understanding of current theoretical developments of Evolutionary Artificial Neural Networks (EANNs) using GAs and other heuristic procedures and how the proposed hybrid and other heuristic procedures can be combined to produce an optimal ANN.

A hybrid of multiobjective Evolutionary Algorithm and HMM-Fuzzy model for time series prediction

In this paper, we introduce a new hybrid of Hidden Markov Model (HMM), Fuzzy Logic and multiobjective Evolutionary Algorithm (EA) for building a fuzzy model to predict non-linear time series data. In this hybrid approach, the HMMs log-likelihood score for each data pattern is used to rank the data and fuzzy rules are generated using the ranked data. We use multiobjective EA to find a range of trade-off solutions between the number of fuzzy rules and the prediction accuracy. The model is tested on a number of benchmark and more recent financial time series data. The experimental results clearly demonstrate that our model is able to generate a reduced number of fuzzy rules with similar (and in some cases better) performance compared with typical data driven fuzzy models reported in the literature.

Automated Analysis of Retinal Vascular Tortuosity on Color Retinal Images

Recent advances in medical imaging modality have enabled us to identify new features in retinal vasculature. One of the features is retinal vascular tortuosity which has been shown to become a predictive factor for cardiovascular diseases and diabetes. The changes in retinal vascular tortuosity might be a sign of severity or improvement of the disease. In this paper, we propose a new method for measuring retinal vascular tortuosity. We adopt a new technique to analyze tortuosity that consider vessel-segment’s width simultaneously. Our proposed method measures vessel-segment’s tortuosity on its edge. A qualitative assessment shows that the method is appropriate for measuring the tortuosity of the vessels in different widths and directions in the image. Finally, a comparison distinguishing tortuous vs. non tortuous vessels demonstrates that the proposed approach may be suitable for predicting or earlier diagnosis of diabetes or cardiovascular diseases.

Vessel Cross-Sectional Diameter Measurement on Color Retinal Image

Vessel cross-sectional diameter is an important feature for analyzing retinal vascular changes. In automated retinal image analysis, the measurement of vascular width is a complex process as most of the vessels are few pixels wide or suffering from lack of contrast. In this paper, we propose a new method to measure the retinal blood vessel diameter which can be used to detect arteriolar narrowing, arteriovenous (AV) nicking, branching coefficients, etc. to diagnose various diseases. The proposed method utilizes the vessel centerline and edge information to measure the width for a vessel cross-section. Using the Adaptive Region Growing (ARG) segmentation technique we obtain the edges of the blood vessels, and then applying the unsupervised texture classification method we segment the blood vessels from where the vessel centerline is obtained. The potential pixels pairs for each centerline pixel are obtained from the edge image that pass through this centerline pixel. We apply a rotational invariant mask to search the pixel pairs from the edge image, and calculate the shortest distance pair which provides the vessel width (or diameter) for that cross-section. The method is evaluated with manually measured width for different vessels’ cross-sectional area. For the automated measurement of vascular width we achieve an average accuracy of 95.8%.

Anti-personnel Mine Detection and Classification Using GPR Image

The automated anti-personnel mine (APM) detection and classification is currently a broad issue. The detection success depends on the feature selection that we obtain from the sensors. Ground penetrating radar (GPR) is one of the established sensors for detecting buried APM. In this paper, we introduce a method which improves the accuracy of detecting APM by using GPR imaging. This method adopts a segmentation technique for feature extraction and neural network as a pattern classifier. A seeded region growing algorithm is applied as region based segmentation for pattern construction following the median filtering and threshold of the original GPR image. A feed forward neural network (FFNN) with backpropagation training is employed for classifying the patterns. The FFNN takes the patterns (APM signature) that are constructed from each salient region and generate the classification. This method significantly improves accuracy in the detection and classification of APM

HMM based Fuzzy Model for Time Series Prediction

This paper presents a hidden Markov model (HMM) based fuzzy rule extraction technique for predicting a time series generated by a chaotic dynamical system. The model uses three sequential phases. Firstly, the HMM is used to partition the input dataset based on the ordering of the calculated log-likelihood values (similarity measures). Then, a recursive top-down algorithm is used to generate the minimum number of rules required to accurately predict the next value in the time series using the training dataset. Finally, a gradient descent method is applied to the extracted fuzzy rules in order to fine-tune the model parameters. The performance of the proposed model is evaluated using a benchmark dataset -the Mackey-Glass time series. The results obtained clearly demonstrate significant improvement in prediction capabilities of the proposed HMM-fuzzy model when compared to the other techniques.

Landmine Feature Extraction and Classification of GPR Data Based on SVM Method

In this paper, the problem of detecting buried landmine is tackled in the feature extraction and classification. Determining the likelihood set of an unknown pattern (feature vector), extracted from ground penetrating radar data by using SVM method. The advantage of SVM method in feature extraction and classification of image processing is: A classifier works well both on the training samples and on previously unseen samples; In addition, the SVM provides, enable a classification performance improvement based on from high feature dimensions to two or three feature dimensions. Finally, SVM method has a standard theory and a good implementation algorithm.