京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第20160号 / 理博第4245号 / 新制||理||1610(附属図書館) / 京都大学大学院理学研究科数学・数理解析専攻 / (主査)教授 望月 新一, 教授 岡本 久, 教授 玉川 安騎男 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

Background: With the advent of the information age, there are many large numbers of services rising which run on several clusters of computers.  Maintaining such large complex systems is a very difficult task. Developers use one tool which is common for almost all software systems, they are the console logs. To troubleshoot problems, developers refer to these logs to solve the issue. Identifying anomalies in the logs would lead us to the cause of the problem, thereby automating the analysis of logs. This study focuses on anomaly detection in logs. Objectives: The main goal of the thesis is to identify different algorithms for anomaly detection in logs, implement the algorithms and compare them by doing an experiment. Methods: A literature review had been conducted for identifying the most suitable algorithms for anomaly detection in logs. An experiment was conducted to compare the algorithms identified in the literature review. The experiment was performed on a dataset of logs generated by Hadoop Data File System (HDFS) servers which consisted of more than 11 million lines of logs. The algorithms that have been compared are K-means, DBSCAN, Isolation Forest, and Local Outlier Factor algorithms which are all unsupervised learning algorithms. Results: The performance of all these algorithms has been compared using metrics precision, recall, accuracy, F1 score, and run time. Though DBSCAN was the fastest, it resulted in poor recall, similarly Isolation Forest also resulted in poor recall. Local Outlier Factor was the fastest to predict. K-means had the highest precision and Local Outlier Factor had the highest recall, accuracy, and F1 score. Conclusion: After comparing the metrics of different algorithms, we conclude that Local Outlier Factor performed better than the other algorithms with respect to most of the metrics measured.

Bottomland hardwood stands of the Mid-South region of the United States are some of the most productive forests in the country. A large percentage of these stands are owned by nonindustrial private forest landowners, who have little information on which to base management decisions. These stands are, therefore, a largely unmanaged and under-utilized reserve of high quality hardwoods. To provide landowners with a decision-making tool for comparing management scenarios, a growth and yield study was initiated in 1981. One hundred and fifty permanent plots were installed in red oaksweetgum stands. The study has been remeasured three times over the past 35 years. New plots were added when losses occurred due to natural disasters or harvesting. Stand level (Iles 2008), log grade volume distribution (Banzhaf 2009), and diameter distribution (Howard 2011) models were developed as component models of the overall growth and yield system. This study completes the modeling effort by developing individual tree equations for percent annual diameter growth and survival. Equations were constructed using linear, non-linear, and logistic regression techniques. The best set of developed equations was selected based on biological consistency, joint behavior when inserted into the growth and yield computer model, and the performance of each plot’s predicted future yield when compared to its observed data at the next projection period. Final independent stand level variables for the two models included age, diameter at breast height, trees per acre, and average height of dominant trees. Percent diameter growth and survival equations exhibited high fit statistics and when coupled with the other equations in the computer model, produced estimates for trees per acre, basal area, arithmetic and quadratic mean diameters with low bias and root mean squared error. The resulting growth and yield simulator implemented in Microsoft Visual Basic® Editor within Microsoft Excel® enables forest professionals and landowners to make better management decisions for their red oak-sweetgum mixture bottomland hardwood stands by projecting current forest inventories into the future, predicting average yields, and evaluating and comparing forest management scenarios.

Bridges and buildings are often supported on deep foundations. These foundations consist of groups of piles coupled together by concrete pile caps. These pile caps, which are often massive and deeply buried, would be expected to provide significant resistance to lateral loads. However, practical procedures for computing the resistance of pile caps to lateral loads have not been developed, and, for this reason, cap resistance is usually ignored. Neglecting cap resistance results in estimates of pile group deflections and bending moments under load that may exceed the actual deflections and bending moments by 100 % or more. Advances could be realized in the design of economical pile-supported foundations, and their behavior more accurately predicted, if the cap resistance can be accurately assessed. This research provides a means of assessing and quantifying many important aspects of pile group and pile cap behavior under lateral loads. The program of work performed in this study includes developing a full-scale field test facility, conducting approximately 30 lateral load tests on pile groups and pile caps, performing laboratory geotechnical tests on natural soils obtained from the site and on imported backfill materials, and performing analytical studies. A detailed literature review was also conducted to assess the current state of practice in the area of laterally loaded pile groups. A method called the "group-equivalent pile" approach (abbreviated GEP) was developed for creating analytical models of pile groups and pile caps that are compatible with established approaches for analyzing single laterally loaded piles. A method for calculating pile cap resistance-deflection curves (p-y curves) was developed during this study, and has been programmed in the spreadsheet called PYCAP. A practical, rational, and systematic procedure was developed for assessing and quantifying the lateral resistance that pile caps provide to pile groups. Comparisons between measured and calculated load-deflection responses indicate that the analytical approach developed in this study is conservative, reasonably accurate, and suitable for use in design. The results of this research are expected to improve the current state of knowledge and practice regarding pile group and pile cap behavior. / Ph. D.

Southern bottomland sites are among the most productive areas for producing high quality grade hardwood, yet the ability to estimate the quantity and quality of standing grade hardwood is almost non-existent. Measurements and observed log grades were recorded on standing trees to construct volume prediction models for individual trees. Several different modeling techniques were explored and compared during development. Developed equations predict merchantable sawtimber volume and volume by grade category in trees by species group. Two separate sets of equations were developed for each species group using either total height or merchantable height. Models were chosen based on significance of variables, index of fit, RMSE, bias, ease of use, and biological trends. The models developed to predict merchantable sawtimber and grade volumes were designed to be implemented in a larger hardwood growth and yield system.

No description available.

No description available.

No description available.

No description available.

No description available.