Machine learning techniques to make computers easier to use

Hiroshi Motoda, Kenichi Yoshida
1998 Artificial Intelligence  
Identifying user-dependent information that can be automatically collected helps build a user model by which (1) to predict what the user wants to do next and (2) to do relevant preprocessing. Such information is often relational and is best represented by a set of directed graphs. A machine learning technique called graph-based induction (GM) efficiently extracts regularities from such data, based on which a user-adaptive interface is built that can predict the next command, generate scripts
more » ... d prefetch files in a multi task environment. The heart of GBI is pairwise chunking. The paper shows how this simple mechanism applies to the top down induction of decision trees for nested attribute representation as well as finding frequently occurring patterns in a graph. The results clearly shows that the dependency analysis of computational processes activated by the user commands which is made possible by GBI is indeed useful to build a behavior model and increase prediction accuracy. 0 1998 Elsevier Science B .V. All rights reserved. user interviews, application-specific heuristics, and stereotypical inferences are often not appropriate, and a better automated method is being sought. Finding regularities in data is a basis of knowledge acquisition, and extracting behavioral patterns from the user information is one such problem. Since each user may do the same thing in a different way, identifying the information that can characterize the user and be automatically collected is crucial. Once such information is found and if an appropriate machine learning technique can induce regularities in each user's behavior to carry out his/her intended task, we can use them to guide the daily work and to do some preprocessing, which may facilitate easiness of usage and increase efficiency. In order for this to work satisfactorily, we rely on the assumption that situation, purpose, intention, meaning, concept are all embedded in some structure, and thus, extractable by mechanical operation. We discuss three learning tasks, command prediction, script generation and file prefetching in a multi task environment. The scope of user behavior is limited to a sequence of task execution (e.g., editing, formatting, viewing, etc.) using plural application programs. Most studies that attempted to develop a user-adaptive interface system only analyzed the sequence of user behaviors, from which to automate the repetitions (see [S]). In this setting, the data can easily be represented by attribute-value pairs, each attribute denoting the sequence order and its value, the command, and a standard classifier, e.g., [22] can be directly applied to induce a set of classification rules without any difficulty. However, since the command sequence does not necessarily typify the user's behavior, the user model constructed from only the sequence information may not adequately capture the user's behavior (we have confirmed this and the results are shown later). We focused on the process I/O information that is also automatically collected along with the command sequence. Since this is dependency information and its relationship cannot be fixed in advance, it is not straightforward to represent this by attribute-value pairs and apply a standard classifier. We show that graph-based induction [25] can nicely be applied to the three learning tasks. In this paper, we revisit GBZ, show how it can extract typical patterns from a set of directed graphs and how it can induce classification rules using a similar technique in the Top Down Decision Tree (TDDT) induction algorithm. The first and the second learning tasks are implemented as ClipBoard which is a window like UNIX shell [26], and the third task is implemented as Prcfetch duemon that is hidden from the user. The results clearly show that the dependency analysis of computational processes activated by the user's commands, which is made possible by GBI, is indeed useful. ClipBoard is in daily use and its prediction accuracy and response time are satisfactory. Prefetch duemon works as expected only for I/O intensive task due to an implementation problem, and thus needs further improvement. The following section introduces the three learning tasks. Subsequent sections describe the learning method GBZ and summarize the results of learning experiments performed to date. The last two sections consider lessons learned from this study and directions for future research.
doi:10.1016/s0004-3702(98)00062-9 fatcat:3egioabdfjbvlpyjlmxazmdoce