Abstract

Abstract

Abstract
Quotations from financial leaders can have significant influence upon the immediate prospects of economic actors. Indiscreet or candid comments from senior business leaders have had detrimental effects upon their organizations. Established polarity classification techniques perform poorly when classifying quotations because they display a number of complex linguistic features and lack of training data. The proposed strategy segments the quotations by inferred "opinion maker" role and then applies individual polarity classification strategies to each group of the segmented quotations. This strategy demonstrates a clear advantage over applying classical classification techniques to the whole corpus of quotations. While modelling contextual information with Random Forests based on a vector of unigrams plus the "opinion maker role" reaches a maximum F-measure of 52.85%, understanding the "bias" of the quotation maker previously based on its lexical usage allows 86.23% F-measure for "unbiased" quotations and 71.10% F-measure for "biased" quotations with the Naive Bayes classifier.

Abstract
News can contain information which may provide an indication of the future direction of a share or stock market index. The possibility of predicting future stock market prices has attracted an increasing numbers of industry practitioners and academic researchers to this area of investigation. Popular approaches have relied upon either: models constructed from manually selected training or manually constructed dictionaries. A potential flaw of manually selecting data is that the effectiveness of the trained model is dependent upon the ability of the human annotator. An alternative approach is to manually align news stories with trends in a specific market. A negative story is inferred if it co-occurs with a market losing value where as positive story is associated with a rise. This approach may have its flaws because news stories may co-occur with market movements by chance and consequently may inhibit the construction of a robust classifier with data gathered by this method. This paper presents a strategy which combines a: rule classifier, alignment strategy and self-training to induce a robust model for classifying news stories. The proposed method is compared with several competing methodologies and is evaluated with: estimated F-Measure and estimated trading returns. In addition the paper provides an evaluation of classifying a news story with its: headline, description or story text with: Language Models and Naive Bayes. The results demonstrate a clear advantage for the proposed methodology when evaluated by estimated F-Measure. The proposed strategy also produces the highest trading returns. In addition the paper clearly demonstrates that a news story's headline provides the greatest assistance for classification. The models induced from headlines gained the highest estimated F-Measure and trading returns for each strategy with the exception of the alignment method which performed uniformly poorly. © Technomathematics Research Foundation.

Abstract
News can contain information which may provide an indication of the future direction of a share or stock market index. The possibility of predicting future stock market prices has attracted an increasing numbers of industry practitioners and academic researchers to this area of investigation. Popular approaches have relied upon either: models constructed from manually selected training or manually constructed dictionaries. A potential flaw of manually selecting data is that the effectiveness of the trained model is dependent upon the ability of the human annotator. An alternative approach is to align news stories with trends in a specific market. A negative story is inferred if it co-occurs with a market losing value where as positive story is associated with a rise. This approach may have its flaws because news stories may co-occur with market movements by chance and consequently may inhibit the construction of a robust classifier with data gathered with this method. This paper presents a strategy which combines a: rule classifier, alignment strategy and self-training to induce a robust model for classifying news stories. The proposed method is compared with several competing methodologies and is evaluated with: estimated F-Measure and estimated trading returns. In addition the paper provides an evaluation of classifying a news story with it's: headline, description or story text. The results demonstrate a clear advantage for the proposed methodology when evaluated by estimated F-Measure. The proposed strategy also produces the highest trading returns. In addition the paper clearly demonstrates that a news story's headline provides the greatest assistance for classification. The models induced from headlines gained the highest estimated F-Measure and trading returns for each strategy with the exception of the alignment method which performed uniformly poorly. © 2011 AISTI.

Abstract
A trading strategy is an algorithm that provides decision support for a trader. An ideal system suggests which stocks to buy and sell at every moment. Limited but still very useful trading strategies suggest stocks to buy, but leave the sell decisions and the decision of proportions of different stocks to the trader, or to another automatic decision mechanisim. In this paper we use a previously introduced method of predicting rank variables to produce both buy and sell decisions. The rank variables are predicted by neural networks, and provide an efficient way to produce daily buy (and also sell) suggestions. This should be seen in contrast to "ordinary" technical indicators that often give very few signals, or buy/sell signals for many stocks at the same time. The produced buy signals are further processed in a classification module that aims at identifying which of the numerous buy signals one should trust, and of which ones one should discard. The classification reduces the number of buy signals and also increases both hitrate and overall profit for a simulated trader. Data from the US stock market for 1992-2001 is used in the tests of the system, and the results show how a trading system's performance can be significantly improved by adding a post-processing classification layer between the generation of trading signals and the actual decision making.