Georg Ruß

Relevance Feedback for Association Rules by Leveraging Concepts from Information Retrieval

The task of detecting those association rules which are interesting within the vast set of discovered ones still is a major research hallenge in data mining. Although several possible solutions have been proposed, they usually require a user to be aware what he knows, to have a rough idea what he is looking for, and to be able to specify this knowledge in advance. In this paper we compare the task of finding the most relevant rules with the task of finding the most relevant documents known from Information Retrieval. We propose a novel and flexible method of relevance feedback for association rules which leverages technologies from Information Retrieval, like document vectors, term frequencies and similarity calculations. By acquiring a user’s preferences our approach builds a repository of what he considers to be (non-)relevant. By calculating and aggregating the similarities of each unexamined rule with the rules in the repository we obtain a relevance score which better reflects the user’s notion of relevance with each feedback provided.

Data Mining of Agricultural Yield Data: A Comparison of Regression Models

Nowadays, precision agriculture refers to the application of state-of-the-art GPS technology in connection with small-scale, sensor-based treatment of the crop. This introduces large amounts of data which are collected and stored for later usage. Making appropriate use of these data often leads to considerable gains in efficiency and therefore economic advantages. However, the amount of data poses a data mining problem --- which should be solved using data mining techniques. One of the tasks that remains to be solved is yield prediction based on available data. From a data mining perspective, this can be formulated and treated as a multi-dimensional regression task. This paper deals with appropriate regression techniques and evaluates four different techniques on selected agriculture data. A recommendation for a certain technique is provided.

Exploratory hierarchical clustering for management zone delineation in precision agriculture

Precision Agriculture has become an emerging topic over the last ten years. It is concerned with the integration of information technology into agricultural processes. This is especially true for the ongoing and growing data collection in agriculture. Novel ground-based sensors, aerial and satellite imagery as well as soil sampling provide large georeferenced data sets with high spatial resolution. However, these data lead to the data mining problem of finding novel and useful information in these data sets. One of the key tasks in the area of precision agriculture is management zone delineation: given a data set of georeferenced data records with high spatial resolution, we would like to discover spatially mostly contiguous zones on the field which exhibit similar characteristics within the zones and different characteristics between zones. From a data mining point of view, this task comes down to a variant of spatial clustering with a constraint of keeping the resulting clusters spatially mostly contiguous. This article presents a novel approach tailored to the specifics of the available data, which do not allow for using an existing algorithm. A variant of hierarchical agglomerative clustering will be presented, in conjunction with a spatial constraint. Results on available multi-variate data sets and subsets will be presented.

Estimation of Neural Network Parameters for Wheat Yield Prediction

Precision agriculture (PA) and information technology (IT) are closely interwoven. The former usually refers to the application of nowadays’ technology to agriculture. Due to the use of sensors and GPS technology, in today’s agriculture many data are collected. Making use of those data via IT often leads to dramatic improvements in efficiency. For this purpose, the challenge is to change these raw data into useful information. This paper deals with suitable modeling techniques for those agricultural data where the objective is to uncover the existing patterns. In particular, the use of feed-forward backpropagation neural networks will be evaluated and suitable parameters will be estimated. In consequence, yield prediction is enabled based on cheaply available site data. Based on this prediction, economic or environmental optimization of, e.g., fertilization can be carried out.

Spatial variable importance assessment for yield prediction in precision agriculture

Precision Agriculture applies state-of-the-art GPS technology in connection with site-specific, sensor-based crop management. It can also be described as a data-driven approach to agriculture, which is strongly connected with a number of data mining problems. One of those is also an inherently important task in agriculture: yield prediction. Given a yield prediction model, which of the predictor variables are the important ones? In the past, a number of approaches have been proposed towards this problem. For yield prediction, a broad variety of regression models for non-spatial data can be adapted for spatial data using a novel spatial cross-validation technique. Since this procedure is at the core of variable importance assessment, it will be briefly introduced here. Given this spatial yield prediction model, a novel approach towards assessing a variable’s importance will be presented. It essentially consists of picking each of the predictor variables, one at a time, permutating its values in the test set and observing the deviation of the model’s RMSE. This article uses two real-world data sets from precision agriculture and evaluates the above procedure.

Feature Selection for Wheat Yield Prediction

Carrying out effective and sustainable agriculture has become an important issue in recent years. Agricultural production has to keep up with an everincreasing population by taking advantage of a field’s heterogeneity. Nowadays, modern technology such as the global positioning system (GPS) and a multitude of developed sensors enable farmers to better measure their fields’ heterogeneities. For this small-scale, precise treatment the term precision agriculture has been coined. However, the large amounts of data that are (literally) harvested during the growing season have to be analysed. In particular, the farmer is interested in knowing whether a newly developed heterogeneity sensor is potentially advantageous or not. Since the sensor data are readily available, this issue should be seen from an artificial intelligence perspective. There it can be treated as a feature selection problem. The additional task of yield prediction can be treated as a multi-dimensional regression problem. This article aims to present an approach towards solving these two practically important problems using artificial intelligence and data mining ideas and methodologies.

Regression models for spatial data: an example from precision agriculture

The term precision agriculture refers to the application of state-of-theart GPS technology in connection with small-scale, sensor-based treatment of the crop. This data-driven approach to agriculture poses a number of data mining problems. One of those is also an obviously important task in agriculture: yield prediction. Given a precise, geographically annotated data set for a certain field, can a seasons yield be predicted? Numerous approaches have been proposed to solving this problem. In the past, classical regression models for non-spatial data have been used, like regression trees, neural networks and support vector machines. However, in a cross-validation learning approach, issues with the assumption of statistical independence of the data records appear. Therefore, the geographical location of data records should clearly be considered while employing a regression model. This paper gives a short overview about the available data, points out the issues with the classical learning approaches and presents a novel spatial cross-validation technique to overcome the problems and solve the aforementioned yield prediction task.

Visualization of Agriculture Data Using Self-Organizing Maps

The importance of carrying out effective and sustainable agriculture is getting more and more obvious. In the past, additional fallow ground could be tilled to raise production. Nevertheless, even in industrialized countries agriculture can still improve on its overall yield. Modern technology, such as GPS-based tractors and sensor-aided fertilization, enables fanners to optimize their use of resources, economically and ecologically. However, these modern technologies create heaps of data that are not as easy to grasp and to evaluate as they have once been. Therefore, techniques or methods are required which use those data to their full capacity — clearly being a data mining task. This paper presents some experimental results on real agriculture data that aid in the first part of the data mining process: understanding and visualizing the data. We present interesting conclusions concerning fertilization strategies which result from data mining.

Spatial data mining in precision agriculture

Technological advances are nowadays often based on improvements in information and data processing capabilities. Even modern agriculture is to a large extent based on adequate data processing, since the usage of novel information devices, GPS-based georeferenced data collection and high-resolution spatial data sets have become standard modes of operation, turning the once uniform site management into site-specific management as one of the most important sub-fields in precision agriculture. On the one hand, the resulting data sets clearly provide the foundations for economic and ecologic improvements. On the other hand, these data sets pose novel challenges for spatial data mining. Two specific tasks are explored in this study: spatial variable importance and management zone delineation. The foundations of this thesis are data originating in site-specific management operations. They typically include electrical conductivity readings, fertilizer applications, soil sampling results, vegetation indicators and yield measurements. These variables are georeferenced, i.e. for a particular point of the site under study the variables and their values are known at a certain spatial resolution. These spatial data sets are furthermore augmented with digital elevation models from which terrain attributes such as slope, wetness index and curvatures are derived. The first of the tasks is concerned with yield prediction and based on an existing dissertation in this area. Yield prediction is handled as a multivariate regression task using spatial data sets. However, taking the spatial relationships of the data sets into account requires some changes in the standard cross-validation to make it aware of spatial relationships in the data sets. Based on this addition, the question can be answered which of a variety of regression models are best suited for yield prediction. Eventually the regression models help to estimate which of the variables are important for yield prediction using permutation-based variable importance measures. The second task is concerned with management zone delineation. Based on a literature review of existing approaches, a lack of exploratory algorithms for this task is concluded, in both the precision agriculture and the computer science domains. Hence, a novel algorithm (HACC-spatial) is developed, fulfilling the requirements posed in the literature. It is based on hierarchical agglomerative clustering incorporating a spatial constraint. The spatial contiguity of the management zones is the key parameter in this approach. Furthermore, hierarchical clustering offers a simple and appealing way to explore the data sets under study, which is one of the main goals of data mining.

Künstliche neuronale Netze

(Kunstliche) neuronale Netze (artificial neural networks) sind informationsverarbeitende Systeme, deren Struktur und Funktionsweise dem Nervensystem und speziell dem Gehirn von Tieren und Menschen nachempfunden sind. Sie bestehen aus einer grosen Anzahl einfacher, parallel arbeitender Einheiten, den sogenannten Neuronen. Diese Neuronen senden sich Informationen in Form von Aktivierungssignalen uber gerichtete Verbindungen zu.