Machine learning for user modeling

At first blush, user modeling appears to be a prime candidate for straightforward application of standard machine learning techniques. Observations of the user's behavior can provide training examples that a machine learning system can use to form a model designed to predict future actions. However, user modeling poses a number of challenges for machine learning that have hindered its application in user modeling, including: the need for large data sets; the need for labeled data; concept drift; and computational complexity. This paper examines each of these issues and reviews approaches to resolving them.

Beacon interacts with cdc2/cdc28-like kinases

Previously we found elevated beacon gene expression in the hypothalamus of obese Psammomys obesus. Beacon administration into the lateral ventricle of P. obesus stimulated food intake and body weight gain. In the current study we used yeast two-hybrid technology to screen for proteins in the human brain that interact with beacon. CLK4, an isoform of cdc2/cdc28-like kinase family of proteins, was identified as a strong interacting partner for beacon. Using active recombinant proteins and a surface plasmon resonance based detection technique, we demonstrated that the three members of this subfamily of kinases (CLK1, 2, and 4) all interact with beacon. Based on the known sequence and functional properties of beacon and CLKs, we speculate that beacon could either modulate the function of key regulatory molecules such as PTP1B or control the expression patterns of specific genes involved in the central regulation of energy metabolism.

Vested interest in the decision to resolve social dilemma conflicts

This exploratory study investigated the degree of correspondence between individual and group interests in the decision to adopt a sanctioning system to manage a shared resource with social dilemma properties. Fifty-two groups of four people accessed a "free-running" computer-simulated shared resource and had either equal or unequal resource access and experienced either an equitable or inequitable sanctioning system. Consistent with past research, a worse group outcome and greater voting for system change was found under the equitable than under the inequitable sanctioning system. However, at the individual level of analysis, the results suggested that the sanctioning systems did not have the same implications for all group members and that some of those differences predicted voting for system change. The study suggests the need to investigate the various decision frames and inequitable implications associated with structural/institutional change and calls for further investigation of dependencies between micro- and macro-level units of analysis.

Proportional k-interval discretization for naive-Bayes classifiers

This paper argues that two commonly-used discretization approaches, fixed k-interval discretization and entropy-based discretization have sub-optimal characteristics for naive-Bayes classification. This analysis leads to a new discretization method, Proportional k-Interval Discretization (PKID), which adjusts the number and size of discretized intervals to the number of training instances, thus seeks an appropriate trade-off between the bias and variance of the probability estimation for naive-Bayes classifiers. We justify PKID in theory, as well as test it on a wide cross-section of datasets. Our experimental results suggest that in comparison to its alternatives, PKID provides naive-Bayes classifiers competitive classification performance for smaller datasets and better classification performance for larger datasets.

Discovering associations with numeric variables

This paper further develops Aumann and Lindell's [3] proposal for a variant of association rules for which the consequent is a numeric variable. It is argued that these rules can discover useful interactions with numeric data that cannot be discovered directly using traditional association rules with discretization. Alternative measures for identifying interesting rules are proposed. Efficient algorithms are presented that enable these rules to be discovered for dense data sets for which application of Auman and Lindell's algorithm is infeasible.

Further pruning for efficient association rule discovery

The Apriori algorithm’s frequent itemset approach has become the standard approach to discovering association rules. However, the computation requirements of the frequent itemset approach are infeasible for dense data and the approach is unable to discover infrequent associations. OPUS AR is an efficient algorithm for association rule discovery that does not utilize frequent itemsets and hence avoids these problems. It can reduce search time by using additional constraints on the search space as well as constraints on itemset frequency. However, the effectiveness of the pruning rules used during search will determine the efficiency of its search. This paper presents and analyses pruning rules for use with OPUS AR. We demonstrate that application of OPUS AR is feasible for a number of datasets for which application of the frequent itemset approach is infeasible and that the new pruning rules can reduce compute time by more than 40%.

The need for low bias algorithms in classification learning from large data sets

This paper reviews the appropriateness for application to large data sets of standard machine learning algorithms, which were mainly developed in the context of small data sets. Sampling and parallelisation have proved useful means for reducing computation time when learning from large data sets. However, such methods assume that algorithms that were designed for use with what are now considered small data sets are also fundamentally suitable for large data sets. It is plausible that optimal learning from large data sets requires a different type of algorithm to optimal learning from small data sets. This paper investigates one respect in which data set size may affect the requirements of a learning algorithm — the bias plus variance decomposition of classification error. Experiments show that learning from large data sets may be more effective when using an algorithm that places greater emphasis on bias management, rather than variance management.