Richard Kyle MacKinnon

Mining constrained frequent itemsets from distributed uncertain data

Nowadays, high volumes of massive data can be generated from various sources (e.g.,sensor data from environmental surveillance). Many existing distributed frequent itemset mining algorithms do not allow users to express the itemsets to be mined according to their intention via the use of constraints. Consequently, these unconstrained mining algorithms can yield numerous itemsets that are not interesting to users. Moreover, due to inherited measurement inaccuracies and/or network latencies, the data are often riddled with uncertainty. These call for both constrained mining and uncertain data mining. In this journal article, we propose a data-intensive computer system for tree-based mining of frequent itemsets that satisfy user-defined constraints from a distributed environment such as a wireless sensor network of uncertain data. We proposed a system for tree-based distributed uncertain frequent itemset mining.Our system allows users to specify constraints for expressing their interests.It finds frequent itemsets that satisfy succinct constraints from distributed uncertain data.It also handles non-succinct (e.g.,inductive succinct, anti-monotone) constraints.

Reducing the Search Space for Big Data Mining for Interesting Patterns from Uncertain Data

Many existing data mining algorithms search interesting patterns from transactional databases of precise data. However, there are situations in which data are uncertain. Items in each transaction of these probabilistic databases of uncertain data are usually associated with existential probabilities, which express the likelihood of these items to be present in the transaction. When compared with mining from precise data, the search space for mining from uncertain data is much larger due to the presence of the existential probabilities. This problem is worsened as we are moving to the era of Big data. Furthermore, in many real-life applications, users may be interested in a tiny portion of this large search space for Big data mining. Without providing opportunities for users to express the interesting patterns to be mined, many existing data mining algorithms return numerous patterns -- out of which only some are interesting. In this paper, we propose an algorithm that (i) allows users to express their interest in terms of constraints and (ii) uses the MapReduce model to mine uncertain Big data for frequent patterns that satisfy the user-specified constraints. By exploiting properties of the constraints, our algorithm greatly reduces the search space for Big data mining of uncertain data, and returns only those patterns that are interesting to the users for Big data analytics.

Fast Algorithms for Frequent Itemset Mining from Uncertain Data

The majority of existing data mining algorithms mine frequent item sets from precise databases. A well-known algorithm is FP-growth, which builds a compact FP-tree structure to capture important contents of the database and mines frequent item sets from the FP-tree. However, there are situations in which data are uncertain. In recent years, researchers have paid attention to frequent item set mining from uncertain databases. UFP-growth is one of the frequently cited algorithms for mining uncertain data. However, the corresponding UFP-tree structure can be large. Other tree structures for handling uncertain data may achieve compactness at the expense of looser upper bounds on expected supports. To solve this problem, we propose two compact tree structures which capture uncertain data with tighter upper bounds than existing tree structures. We also designed two algorithms that mine frequent item sets from our proposed trees. Our experimental results show the tightness of bounds to expected supports provided by these algorithms.

Finding groups of friends who are significant across multiple domains in social networks

Social networking websites such as Facebook, LinkedIn, Twitter, and Weibo have been used for collaboration and knowledge sharing between users. The mining of social network data has become an important topic in data mining and computational aspects of social networks. Nowadays, it is not uncommon for most users in a social network to have many friends and in multiple social domains. Among these friends, some groups of friends are more significant than others. In this paper, we introduce a data mining technique that helps social network users find groups of friends who are significant across multiple domains in social networks.

BLIMP: A Compact Tree Structure for Uncertain Frequent Pattern Mining

Tree structures (e.g., UF-trees, UFP-trees) corresponding to many existing uncertain frequent pattern mining algorithms can be large. Other tree structures for handling uncertain data may achieve compactness at the expense of loose upper bounds on expected supports. To solve this problem, we propose a compact tree structure that captures uncertain data with tighter upper bounds than the aforementioned tree structures. The corresponding algorithm mines frequent patterns from this compact tree structure. Experimental results show the compactness of our tree structure and the tightness of upper bounds to expected supports provided by our uncertain frequent pattern mining algorithm.

Stock Price Prediction in Undirected Graphs Using a Structural Support Vector Machine

Business analytics techniques help mine and analyze business/financial data. For instance, a structural support vector machine (SSVM) can be used to perform classification on complex inputs such as the nodes of a graph structure. We connect collaborating companies in the information technology sector in an undirected graph and use an SSVM to predict positive or negative movement in their stock prices. By using a minimum graph-cutting algorithm to drive the cutting plane optimization problem of the SSVM, an exact solution is achieved in polynomial time. The learned model exploits the associative relationship between the prices of the collaborating companies to outperform the accuracy of a regular SVM. Experiments were conducted using the companies in the Standard and Poors 500-45 Information Technology Sector index. Trades based on the learned model achieved superior returns in the range of 10% to 17% while tracking the index alone over the same time periods yielded returns in the range of -17% to 9%.

Tightening Upper Bounds to the Expected Support for Uncertain Frequent Pattern Mining

Abstract Due to advances in technology, high volumes of valuable data can be collected and transmitted at high velocity in various scientific and engineering applications. Consequently, efficient data mining algorithms are in demand for analyzing these data. For instance, frequent pattern mining discovers implicit, previously unknown, and potentially useful knowledge about relationships among frequently co-occurring items, objects and/or events. While many frequent pattern mining algorithms handle precise data, there are situations in which data are uncertain . In recent years, tree-based algorithms for mining uncertain data have been developed. However, tree structures corresponding to these algorithms can be large. Other tree structures for handling uncertain data may achieve compactness at the expense of loose upper bounds on expected supports. In this paper, we propose (i) a compact tree structure for capturing uncertain data, (ii) a technique for using our tree structure to tighten upper bounds to expected support, and (iii) an algorithm for mining frequent patterns based on our tightened bounds. Experimental results show the benefits of our tightened upper bounds to expected supports in uncertain frequent pattern mining.

BigSAM: Mining Interesting Patterns from Probabilistic Databases of Uncertain Big Data

Nowadays, high volumes of valuable uncertain data can be easily collected or generated at high velocity in many real-life applications. Mining these uncertain Big data is computationally intensive due to the presence of existential probability values associated with items in every transaction in the uncertain data. Each existential probability value expresses the likelihood of that item to be present in a particular transaction in the Big data. In some situations, users may be interested in mining all frequent patterns from these uncertain Big data; in other situations, users may be interested in only a tiny portion of these mined patterns. To reduce the computation and to focus the mining for the latter situations, we propose a tree-based algorithm that (i) allows users to express the patterns to be mined according to their intention via the use of constraints and (ii) uses MapReduce to mine uncertain Big data for only those frequent patterns that satisfy user-specified constraints. Experimental results show the effectiveness of our algorithm in mining probabilistic databases of uncertain Big data.

Finding efficiencies in frequent pattern mining from big uncertain data

Many existing data mining algorithms search interesting patterns from transactional databases of precise data. However, there are situations in which data are uncertain. Items in each transaction of these probabilistic databases of uncertain data are usually associated with existential probabilities, which express the likelihood of these items to be present in the transaction. When compared with mining from precise data, the search space for mining from uncertain data is much larger due to the presence of the existential probabilities. This problem is worsened as we are moving to the era of Big data. Furthermore, in many real-life applications, users may be interested in a tiny portion of this large search space for Big data mining. Without providing opportunities for users to express the interesting patterns to be mined, many existing data mining algorithms return numerous patterns—out of which only some are interesting. In this article, we propose an algorithm that allows users to express their interest in terms of constraints, uses the MapReduce model to mine uncertain Big data for frequent patterns that satisfy the user-specified anti-monotone and monotone constraints, as well as balance the load.

A machine learning approach for stock price prediction

Data mining and machine learning approaches can be incorporated into business intelligence (BI) systems to help users for decision support in many real-life applications. Here, in this paper, we propose a machine learning approach for BI applications. Specifically, we apply structural support vector machines (SSVMs) to perform classification on complex inputs such as the nodes of a graph structure. We connect collaborating companies in the information technology sector in a graph structure and use an SSVM to predict positive or negative movement in their stock prices. The complexity of the SSVM cutting plane optimization problem is determined by the complexity of the separation oracle. It is shown that (i) the separation oracle performs a task equivalent to maximum a posteriori (MAP) inference and (ii) a minimum graph cutting algorithm can solve this problem in the stock price case in polynomial time. Experimental results show the practicability of our proposed machine learning approach in predicting stock prices.