Bibliographic record
Abstract
Online analytical processing (OLAP) and data mining are two important data analysis methods. To improve the functionality of OLAP systems, summary mining aims to find interesting summaries for a data set. In this thesis, we introduce a summary mining architecture, called GenSpace summary mining (GSSM), based on belief revision. Under this framework, a user's beliefs are represented by the user's estimated probability distribution (or estimates) for records in summaries. The summaries are organized in a graphical structure called a GenSpace graph. The GenSpace graph contains information about the conceptual levels of the summaries, the observed probability distributions of records in the summaries, and the user's estimates. The interestingness of the summaries is defined as the distance between the user's estimates and the observed probability distributions. During the mining process, the user specifies his/her estimates at a certain conceptual level in a GenSpace, and the system propagates them to other conceptual levels. The interesting summaries, i.e., the summaries far from the user's estimates, are then selected for presentation to the user. With the interesting summaries as new evidence, the user can revise his/her estimates and input them into the system to start the next round of the mining process. The GSSM process is iterative and it can be interactive. GenSpace summary mining consists of two parts, GenSpace estimate propagation (GSEP) and GenSpace summary selection (GSSS). The GSEP process is a probability based belief revision process. We argue that GSEP should preserve the consistency of the estimates in the GenSpace graph, be efficient, and guarantee the minimum change to the old estimates. Based on these principles, we formalize the GSEP problem as an optimization problem and propose a linear GSEP method as a heuristic approach to solving this problem. We then propose two pruning and path selection strategies to improve the propagation efficiency in GenSpace subgraphs. We also introduce virtual bottom nodes to further reduce the propagation and storage costs during the GSEP process. The experiments indicate that these techniques can greatly improve propagation efficiency. For the GSSS process, we study nine interestingness measures for summaries and their properties. These properties can be used as pruning strategies during the summary selection process to improve system efficiency. We demonstrate the effectiveness of the GSSM method on three real data sets, the Saskatchewan weather data set, the University of Regina student data set, and a customer data set. Experiments in all three data sets show that when the user accepts the distribution of the most interesting summary, its closely related summaries become less interesting in the following mining round and the user's estimates of all summaries approach the observed distributions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".