Advances in Information Mining

44-51

Authors: RAJKUMAR, SAKHARE D.Y

In recent times, the necessity of generating single document summary has gained popularity among the researchers due to its extensive applicability. Most of the automatic text summarization systems utilize extraction-based techniques for selecting the most significant portions of text to generate coherent summaries. In this paper we will analyze the performance of fuzzified neural network approach with the graph theory approach. In the proposed system, we have developed an efficient automatic text summarization system based neural network and fuzzy logic. In the training phase at first, the feature vector is computed for a set of sentences using the feature extraction technique. After that, the feature vector and their corresponding fuzzy score are used to train the neural network optimally. Later in the testing phase, the input document is subjected to preprocessing and feature extraction techniques. In order to obtain the sentence score for every sentence in the input document, the feature vector is fed to the trained neural network that returns the sentence score for every sentence. Finally, by making use of sentence score, the most important sentences are extracted from the input document. The experimentation is performed with the DUC 2002 dataset and the generated summary is evaluated with the measures such as Precision, recall and f-measure. The comparative results of our proposed approach with the graph theory approach produces better results by means of different compression rates.

57-66

Authors: SAIGAL S., MEHROTRA D

This paper focuses on the methodology used in applying the Time Series Data Mining techniques to financial time series data for calculating currency exchange rates of US dollars to Indian Rupees. Four Models namely Multiple Regression in Excel, Multiple Linear Regression of Dedicated Time Series Analysis in Weka, Vector Autoregressive Model in R and Neural Network Model using NeuralWorks Predict are analyzed. All the models are compared on the basis of the forecasting errors generated by them. Mean Error (ME), Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Square Error (RMSE), Mean Percentage Error (MPE) and Mean Absolute Percentage Error (MAPE) are used as a forecast accuracy measure. Results show that all the models accurately predict the exchange rates, but Multiple Linear Regression of Dedicated Time Series Analysis in Weka outperforms the other three models.