In this thesis is presented design and implementation of a system performing genetic programming in simply typed lambda calculus. Population initialization method based on term generating technique producing typed lambda terms in long normal form is introduced. This method is parameterized by simple search strategy. Several search strategies are presented, such as strategy for systematic generation or strategy corresponding to standard ramped half-and-half method. Another such a strategies called \textit{geometric} strategy is further examined in experiments and shown to have various desirable effects such as improved success rate, lesser time consumption and smaller average term size in comparison with standard ramped half-and-half generating method. Other performance enhancements are proposed and supported by experiments such as eta-normalization of generated individuals and @-tree representation of individuals. Abstraction elimination is utilized to enable use of simple tree- swapping crossover. Powered by TCPDF (www.tcpdf.org)

Sin Man Cheang. / "March 2005." / Thesis (Ph.D.)--Chinese University of Hong Kong, 2005. / Includes bibliographical references (p. 219-233) / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.

Genetic programming is an approach that utilises the power of evolution to allow computers to evolve programs. While loops are natural components of most programming languages and appear in every reasonably-sized application, they are rarely used in genetic programming. The work is to investigate a number of restricted looping constructs to determine whether any significant benefits can be obtained in genetic programming. Possible benefits include: Solving problems which cannot be solved without loops, evolving smaller sized solutions which can be more easily understood by human programmers and solving existing problems quicker by using fewer evaluations. In this thesis, a number of explicit restricted loop formats were formulated and tested on the Santa Fe ant problem, a modified ant problem, a sorting problem, a visit-every-square problem and a difficult object classification problem. The experimental results showed that these explicit loops can be success fully used in genetic programming. The evolutionary process can decide when, where and how to use them. Runs with these loops tended to generate smaller sized solutions in fewer evaluations. Solutions with loops were found to some problems that could not be solved without loops. The results and analysis of this thesis have established that there are significant benefits in using loops in genetic programming. Restricted loops can avoid the difficulties of evolving consistent programs and the infinite iterations problem. Researchers and other users of genetic programming should not be afraid of loops.

The domain of medical imaging analysis has burgeoned in recent years due to the availability and affordability of digital radiographic imaging equipment and associated algorithms and, as such, there has been significant activity in the automation of the medical diagnostic process. One such process, cephalometric analysis, is manually intensive and it can take an experienced orthodontist thirty minutes to analyse one radiology image. This thesis describes an approach, based on genetic programming, neural networks and machine learning, to automate this process. A cephalometric analysis involves locating a number of points in an X-ray and determining the linear and angular relationships between them. If the points can be located accurately enough, the rest of the analysis is straightforward. The investigative steps undertaken were as follows: Firstly, a previously published method, which was claimed to be domain independent, was implemented and tested on a selection of landmarks, ranging from easy to very difficult. These included the menton, upper lip, incisal upper incisor, nose tip and sella landmarks. The method used pixel values, and pixel statistics (mean and standard deviation) of pre-determined regions as inputs to a genetic programming detector. This approach proved unsatisfactory and the second part of the investigation focused on alternative handcrafted features sets and fitness measures. This proved to be much more successful and the third part of the investigation involved using pulse coupled neural networks to replace the handcrafted features with learned ones. The fourth and final stage involved an analysis of the evolved programs to determine whether reasonable algorithms had been evolved and not just random artefacts learnt from the training images. A significant finding from the investigative steps was that the new domain independent approach, using pulse coupled neural networks and genetic programming to evolve programs, was as good as or even better than one using the handcrafted features. The advantage of this finding is that little domain knowledge is required, thus obviating the requirement to manually generate handcrafted features. The investigation revealed that some of the easy landmarks could be found with 100\% accuracy while the accuracy of finding the most difficult ones was around 78\%. An extensive analysis of evolved programs revealed underlying regularities that were captured during the evolutionary process. Even though the evolutionary process took different routes and a diverse range of programs was evolved, many of the programs with an acceptable detection rate implemented algorithms with similar characteristics. The major outcome of this work is that the method described in this thesis could be used as the basis of an automated system. The orthodontist would be required to manually correct a few errors before completing the analysis.

Several studies indicate that the data-driven models have proven to be potentially useful tools in hydrological modeling. Nevertheless, it is a common perception among researchers and practitioners that the usefulness of the system theoretic models is limited to forecast applications, and they cannot be used as a tool for scientific investigations. Also, the system-theoretic models are believed to be less reliable as they characterize the hydrological processes by learning the input-output patterns embedded in the dataset and not based on strong physical understanding of the system. It is imperative that the above concerns needs to be addressed before the data-driven models can gain wider acceptability by researchers and practitioners.<p>In this research different methods and tools that can be adopted to promote transparency in the data-driven models are probed with the objective of extending the usefulness of data-driven models beyond forecast applications as a tools for scientific investigations, by providing additional insights into the underlying input-output patterns based on which the data-driven models arrive at a decision. In this regard, the utility of self-organizing networks (competitive learning and self-organizing maps) in learning the patterns in the input space is evaluated by developing a novel neural network model called the spiking modular neural networks (SMNNs). The performance of the SMNNs is evaluated based on its ability to characterize streamflows and actual evapotranspiration process. Also the utility of self-organizing algorithms, namely genetic programming (GP), is evaluated with regards to its ability to promote transparency in data-driven models. The robustness of the GP to evolve its own model structure with relevant parameters is illustrated by applying GP to characterize the actual-evapotranspiration process. The results from this research indicate that self-organization in learning, both in terms of self-organizing networks and self-organizing algorithms, could be adopted to promote transparency in data-driven models.<p>In pursuit of improving the reliability of the data-driven models, different methods for incorporating uncertainty estimates as part of the data-driven model building exercise is evaluated in this research. The local-scale models are shown to be more reliable than the global-scale models in characterizing the saturated hydraulic conductivity of soils. In addition, in this research, the importance of model structure uncertainty in geophysical modeling is emphasized by developing a framework to account for the model structure uncertainty in geophysical modeling. The contribution of the model structure uncertainty to the predictive uncertainty of the model is shown to be larger than the uncertainty associated with the model parameters. Also it has been demonstrated that increasing the model complexity may lead to a better fit of the function, but at the cost of an increasing level of uncertainty. It is recommended that the effect of model structure uncertainty should be considered for developing reliable hydrological models.

Ultrasonography is widely used in medical diagnosis with the advantages of being low cost, non-invasive and capable of real time imaging. When interpreting ultrasonographic images of mammalian ovaries, the structures of interest are follicles, corpora lutea (CL) and stroma. This thesis presents an approach to perform CL texture analysis, including detection and segmentation, based on the classiers trained by genetic programming (GP). The objective of CL detection is to determine whether there is a CL in the ovarian images, while the goal of segmentation is to localize the CL within the image. Genetic programming (GP) oers a solution through the evolution of computer programs by methods inspired by the mechanisms of natural selection. Herein, we use rotationally invariant local binary patterns (LBP) to encode the local texture features. These are used by the programs which are manipulated by GP to obtain highly t CL classiers. Grayscale standardization was performed on all images in our data set based on the reference grayscale in each image. CL classication programs were evolved by genetic programming and tested on ultrasonographic images of ovaries. On the bovine dataset, our CL detection algorithm is reliable and robust. The detection algorithm correctly determined the presence or absence of a CL in 93.3% of 60 test images. The segmentation algorithm achieved a mean ( standard deviation) sensitivity and specicity of 0.87 (0.14) and 0.91 (0.05), respectively, over the 30 CL images. Our CL segmentation algorithm is an improvement over the only previously published algorithm, since our method is fully automatic and does not require the placement of an initial contour. The success of these algorithms demonstrates that similar algorithms designed for analysis of in vivo human ovaries are likely viable.

Ultrasonography is widely used in medical diagnosis with the advantages of being low cost, non-invasive and capable of real time imaging. When interpreting ultrasonographic images of mammalian ovaries, the structures of interest are follicles, corpora lutea (CL) and stroma. This thesis presents an approach to perform CL texture analysis, including detection and segmentation, based on the classiers trained by genetic programming (GP). The objective of CL detection is to determine whether there is a CL in the ovarian images, while the goal of segmentation is to localize the CL within the image. Genetic programming (GP) oers a solution through the evolution of computer programs by methods inspired by the mechanisms of natural selection. Herein, we use rotationally invariant local binary patterns (LBP) to encode the local texture features. These are used by the programs which are manipulated by GP to obtain highly t CL classiers. Grayscale standardization was performed on all images in our data set based on the reference grayscale in each image. CL classication programs were evolved by genetic programming and tested on ultrasonographic images of ovaries. On the bovine dataset, our CL detection algorithm is reliable and robust. The detection algorithm correctly determined the presence or absence of a CL in 93.3% of 60 test images. The segmentation algorithm achieved a mean ( standard deviation) sensitivity and specicity of 0.87 (0.14) and 0.91 (0.05), respectively, over the 30 CL images. Our CL segmentation algorithm is an improvement over the only previously published algorithm, since our method is fully automatic and does not require the placement of an initial contour. The success of these algorithms demonstrates that similar algorithms designed for analysis of in vivo human ovaries are likely viable.

Several studies indicate that the data-driven models have proven to be potentially useful tools in hydrological modeling. Nevertheless, it is a common perception among researchers and practitioners that the usefulness of the system theoretic models is limited to forecast applications, and they cannot be used as a tool for scientific investigations. Also, the system-theoretic models are believed to be less reliable as they characterize the hydrological processes by learning the input-output patterns embedded in the dataset and not based on strong physical understanding of the system. It is imperative that the above concerns needs to be addressed before the data-driven models can gain wider acceptability by researchers and practitioners.<p>In this research different methods and tools that can be adopted to promote transparency in the data-driven models are probed with the objective of extending the usefulness of data-driven models beyond forecast applications as a tools for scientific investigations, by providing additional insights into the underlying input-output patterns based on which the data-driven models arrive at a decision. In this regard, the utility of self-organizing networks (competitive learning and self-organizing maps) in learning the patterns in the input space is evaluated by developing a novel neural network model called the spiking modular neural networks (SMNNs). The performance of the SMNNs is evaluated based on its ability to characterize streamflows and actual evapotranspiration process. Also the utility of self-organizing algorithms, namely genetic programming (GP), is evaluated with regards to its ability to promote transparency in data-driven models. The robustness of the GP to evolve its own model structure with relevant parameters is illustrated by applying GP to characterize the actual-evapotranspiration process. The results from this research indicate that self-organization in learning, both in terms of self-organizing networks and self-organizing algorithms, could be adopted to promote transparency in data-driven models.<p>In pursuit of improving the reliability of the data-driven models, different methods for incorporating uncertainty estimates as part of the data-driven model building exercise is evaluated in this research. The local-scale models are shown to be more reliable than the global-scale models in characterizing the saturated hydraulic conductivity of soils. In addition, in this research, the importance of model structure uncertainty in geophysical modeling is emphasized by developing a framework to account for the model structure uncertainty in geophysical modeling. The contribution of the model structure uncertainty to the predictive uncertainty of the model is shown to be larger than the uncertainty associated with the model parameters. Also it has been demonstrated that increasing the model complexity may lead to a better fit of the function, but at the cost of an increasing level of uncertainty. It is recommended that the effect of model structure uncertainty should be considered for developing reliable hydrological models.

In this thesis, we propose two genetic-programming-based models that improve the trading strategies for mutual funds. These two models can help investors get returns and reduce risks. The first model increases the return by selecting funds with high Sortino ratios and allocates the capital equally, achieving the best annualized return. The second model also selects funds with high Sortino ratios, but reduces the risk by allocating the capital with the mean variance model. Most importantly, our model utilizes the genetic programming to generate feasible trading strategies to gain return, which is suitable for the market that changes anytime. To verify our models, we simulate the investment for mutual funds from January 1999 to December 2009 (11 years in total). The experimental results show that our first model can gain return from 2004/1/1 to 2008/12/31, achieving the best annualized return 9.11%, which is better than the annualized return 6.89% of previous approaches. In addition, our second model with smaller downside volatility can achieve almost the same return as previous results.

Thesis (M.S.)--Worcester Polytechnic Institute. / Keywords: comprehensiblity; technical analysis; genetic programming; overfitting; cooperative coevolution. Includes bibliographical references (p. 82-87).