Ant-Miner Algorithm Research Articles

An Analysis on Qualitative Bankruptcy Prediction Rules using Ant-Miner

Qualitative bankruptcy prediction rules represent experts' problem-solving knowledge to predict qualitative bankruptcy. The objective of this research is predicting qualitative bankruptcy using ant- miner algorithm. Qualitative data are subjective and more difficult to measure. This approach uses qualitative risk factors which include fourteen internal risk factors and sixty eight external risk factors associated with it. By using these factors qualitative prediction rules are generated using ant-miner algorithm and the influence of these factors in bankruptcy is also analyzed. Ant-Miner algorithm is a application of ant colony optimization and data mining concepts. Qualitative rules generated by ant miner algorithm are validated using measure of agreement. These prediction rules yields better accuracy with lesser number of terms than previously applied qualitative bankruptcy prediction methodologies.

Ant Colony Algorithms for Data Learning

Fuzzy Ant-Miner algorithm processes data with nominal class and has the disadvantage of not treating the data with continuous class. In this paper, after presenting the Fuzzy Ant Miner algorithm, the authors propose a new learning method to partition heterogeneous data with continuous class. This method in a first step finds the optimal path between the data using algorithms of ants. Distance adopted in their optimization method takes into account all types of data. The second step vise to divide the data into homogeneous groups by browsing the optimal path found. A new test probability is estimated based on the distance and the amount of pheromone deposited by ants in the transitions between the data. A third step is to find the prototype of each cluster to identify the cluster membership of any new data injected.

Optical Character Recognition using Ant Miner Algorithm: A Case Study on Oriya Character Recognition

Optical Character Recognition (OCR) is one of the challenging areas in the domain of image processing, where the handwritten or printed characters are digitized by using an optical scanner. The image is then analyzed broadly by two methods – (i) matrix space analysis method and (ii) feature space analysis method. Matrix space analysis method takes more memory space and time, compared to feature space analysis. However, it works fine for the scripts in which the strokes are prominent, e.g. English numeric scripts. On the other hand, the feature analysis method is useful where the scripts are complex and having more similarity between the letters in its writing style. Hence, the feature analysis approach is more useful to many of the regional languages. In this paper, we have used the Ant-miner algorithm (AMA) for offline OCR of hand written Oriya scripts, popularly known as Utkal lipi. The AMA is a rule-based approach. The rules are incrementally tuned during the training. The Oriya language contains more than 50 distinct characters i.e. 12 Swara-varnas (i.e., vowels) and 38 Byanjan-varnas (i.e., consonants) and their composite characters. In this work, for the analysis, we define three types of ‘block’s as per the writing styles of the scripts. AMA is then tested with four characters from each ‘block’. Finally, a character recognition tool has been developed using Matlab for observation and validation. General terms: Ant miner algorithm, Optical character recognition (OCR), Oriya character pattern matching

Simulating the spread of opinions in online social networks when targeting opinion leaders

An increasing number of people are joining online social networks. By interacting with each other, network members influence one another's opinion. These influencing effects can be utilized by marketing. A wave of influence can be triggered by addressing only a few opinion leaders in the network. Targeting the right opinion leaders is a big challenge. This paper presents a new approach which simulates the spread of opinions when influencing certain opinion leaders. In contrast to other approaches, the influencing effects are not assumed but revealed by real data. The principles of opinion formation are detected by ant mining algorithms before they are applied to simulate the spread of opinions. The approach is applied to an online gaming community and provides valuable insights for marketing.

Comparative Study on Bio-inspired Approach for Soil Classification

miner is a data mining algorithm based on Ant Colony Optimization. Ant miner algorithms are mainly for discovery rule for optimization. Ant miner+ algorithm uses MAX-MIN ant system for discover rules in the database. Soil classification deals with the systematic categorization of soils based on distinguished characteristics as well as criteria. In this paper, Ant miner and Ant miner+ algorithm were applied to both training and soil dataset to obtain classification rules and found that Ant miner+ performs better than Ant miner.

Mining classification rules for gene expression data using enhanced quickreduct fuzzy-rough feature selection and ant colony optimisation

In this paper, the mining classification rules for gene expression data of yeast S.Cerevisiae is carried out using enhanced quickreduct fuzzy-rough feature selection (EQRFR-FS) with ant colony optimisation (ACO). Rough sets theory deals with uncertainty and vagueness of an information system in data mining. The ACO algorithms have been applied to combinatorial optimisation problems and data mining classification problems. The analyses are carried out in two phases. In the first phase, the features are extracted from the gene expression data and then feature selection is carried out from the extracted features using EQRFR-FS algorithm. In the second phase, from reduct set the classification rules mining are carried out with ant-miner algorithm. The dataset is obtained from the expression profiles maintained by Brown’s group at Stanford University.

A correlation-based ant miner for classification rule discovery

In recent years, a few sequential covering algorithms for classification rule discovery based on the ant colony optimization meta-heuristic (ACO) have been proposed. This paper proposes a new ACO-based classification algorithm called AntMiner-C. Its main feature is a heuristic function based on the correlation among the attributes. Other highlights include the manner in which class labels are assigned to the rules prior to their discovery, a strategy for dynamically stopping the addition of terms in a rule’s antecedent part, and a strategy for pruning redundant rules from the rule set. We study the performance of our proposed approach for twelve commonly used data sets and compare it with the original AntMiner algorithm, decision tree builder C4.5, Ripper, logistic regression technique, and a SVM. Experimental results show that the accuracy rate obtained by AntMiner-C is better than that of the compared algorithms. However, the average number of rules and average terms per rule are higher.

COMPATIBILITY AS A HEURISTIC FOR CONSTRUCTION OF RULES BY ARTIFICIAL ANTS

Ant Miner algorithms are used for the classification task of data mining. This paper proposes a new version of Ant Miner algorithm called CAnt Miner that considers compatibility between the attributes, when constructing a rule. In addition to a new heuristic function based on compatibility among attributes, we also introduce a strategy to dynamically stop adding terms in the rule antecedent part instead of using a user specified threshold that specifies when to stop adding terms. We compare the performance of proposed approach with other Ant Miner versions and C4.5 in nine public domain data set. The experiment results show that the proposed approach achieves higher accuracy rate then other approaches.

Discovering Unordered Rule Sets for Mixed Variables Using an Ant-Miner Algorithm

This work proposes a data mining algorithm called Unordered Rule Sets using a continuous Ant-Miner algorithm. The goal of this work is to extract classification rules from data. Swarm intelligence (SI) is a technique whereby rules may be discovered through the study of collective behavior in decentralized, self-organized systems, such as ants. The Ant-Miner algorithm, first proposed by Parpinelli and his colleagues (2002), applies an ant colony optimization (ACO) heuristic to the classification task of data mining to discover an ordered list of classification rules. Ant-Miner is a rule-induction algorithm that uses SI techniques to form rules. Ant-Miner uses a discretization process to deal with continuous attributes in the data. Discretization transforms numeric attributes into nominal attributes. Discretization may suffer from a loss of information, as the real relationship underlying individual values of a numeric attribute is unknown. The objective of this work is to apply ACO heuristic techniques to discover unordered rule sets for mixed variables in a data set. The proposed algorithm handles both nominal and continuous attributes using multimodal functions. It has the advantage of discovering more modular rules, i.e., rules that can be interpreted independently from other rules - unlike the rules in an ordered list, where the interpretation of a rule requires knowledge of the previous rules in the list. The results provide evidence that the accuracy of the Unordered Rule Set Continuous Ant-Miner algorithm is competitive with other Ant-Miner versions and generates simpler rule sets.