China
Africa
Explore chapters and articles related to this topic
Freshwater Sediment Quality Criteria: Toxicity Bioassessment
Published in Renato Baudo, John P. Giesy, Herbert Muntau, Sediments:, 2020
Methods of data analysis may include various types of hypothesis testing such as Mests or analysis of variance, probit analysis, or various types of multivariate analyses. Determination of the comparative effects of contaminated sediments generally involves statistical evaluation of the null hypothesis that no difference exists between the mean of the response variable in sediments from different locations. Power analysis for the optimization of sample replication is an integral component of the experimental design. To maximize the utility of the results of bioassays, one should determine the statistical power of the bioassay. Specifically, the sources of variation during a test should be identified and unexplained within- and among-test variation minimized (Giesy & Allred 1985). The degree of resolution required for a particular test must also be defined. Once these estimates and decisions have been made, the power of the test can be optimized by selecting the required replication to demonstrate some defined difference between the responses in two or more tests (Giesy & Hoke 1989).
Updates on electrospinning process models (part i)
Published in A. K. Haghi, Lionello Pogliani, Francisco Torrens, Devrim Balköse, Omari V. Mukbaniani, Andrew G. Mercader, Applied Chemistry and Chemical Engineering, 2017
Shima. Maghsoodlou, S. Poreskandar
Over the past few decades, rapid advancements in computer technologies have allowed organizations to gather vast quantities of data for various purposes. Database technology provides efficient access to a myriad of information and allows the creation and maintenance of massive databases. However, it is far easier to collect data than to analyze it and extract information from it. Traditionally, data were analyzed manually but as it grows in size, many hidden patterns and potentially useful relationships may not be recognized by the researcher. As data-manipulation technologies rapidly advance, people rely increasingly on computers to accumulate data, process data, and make use of data. Knowledge discovery in databases consists of intelligent tools that handle massive data sets and useful patterns that help people make use of the data. Also, accurate classification is important for making effective decisions. Accuracy of the results produced by a classification system greatly depends on how well a problem is represented using a set of features. This resulted in the growing interest in the field of Knowledge Discovery in Databases, and a particular process of it, data mining. Data mining applies data analysis and discovery algorithms to identify patterns in data.15
Research Methodology
Published in Zakari Mustapha, Clinton Aigbavboa, Wellington Thwala, Contractor Health and Safety Compliance for Small to Medium-Sized Construction Companies, 2017
Zakari Mustapha, Clinton Aigbavboa, Wellington Thwala
Coding the responses, data cleaning, screening the data and selecting the appropriate data analysis strategy are steps involved in data analysis. Coding of the questionnaire involved identifying, classifying and assigning a numeric or character symbol to data, which may be done in two ways: pre-coded and post-coded (Wong, 1999). The aspect of data analysis from the questionnaire survey will be discussed later in the book. Taken from the list of responses, a number corresponding to a particular selection was given. This process was applied to every question that needed this treatment. Upon completion, the data was then entered into a statistical analysis software package (SPSS) for the next analysis steps. In choosing the appropriate statistical analysis technique, the research elements were considered, namely the research problem, objectives, characteristics of data and the underlying properties of the statistical techniques (Malhotra, 1999). To meet the purposes of this study, descriptive and inferential analyses, and the goodness-of-fit measures of the model were applied where necessary. The data analysis involved the use of multiple analytical techniques to facilitate ease of communicating the results, while at the same time improving its validity. Hence, the use of SEM utilizing EQS software. Raw data from the questionnaire were entered into the SPSS software and were later exported to the SEM software EQS, version 6.2, for analysis. The motivation for the choice of the SEM and particularly the use of the software EQS is explained in the next section.
The relevance of tea culture and recreation industry development based on intelligent big data analysis
Published in Production Planning & Control, 2023
The tea culture tourism planning of tea gardens is the design, layout and arrangement of a series of protection, utilisation, development, etc., based on the current situation and characteristics of the tea culture tourism resources of the tea gardens. Data mining, a prominent approach in big data analysis, identifies similarities from massive data sets using a combination of statistical and machine learning techniques. An illustration would be the analysis of consumer information to establish the groups most inclined to respond to an offering. Tourism planning should be freed from the macro planning of resources and development goals, and the tourism elements should be expressed in the form of planning intention maps, which is a concrete representation of the planning and design process. The stages of planning objectives are creation of goals, establishing a schedule, making the concept official, selectin a monitoring and evaluation methodology, identifying the assets required to carry out those activities, and so on. In general, highly involved, accessibility, as well as engagement define macro-tourism. A macroeconomic strategy is a program’s extremely high-level overview. With no needless filler and no up-to-the-minute planning, it is brief and only hits the key points—what you actually need to understand to do the task.
Behavior Analysis with Machine Learning Using R,
Published in Technometrics, 2022
Chapter 1, “Introduction to Behavior and Machine Learning,” defines behavior as a human or animal observable activity in a particular situation, which can be documented/archived for future analysis. Machine learning (ML) is described as a set of algorithms for automatic discovering of patterns/relationships in data by predictive models working with classifiers and regressors. Different types of ML are considered, such as supervised, semi- supervised, partially- supervised, unsupervised, and reinforcement learning (RL). Data analysis pipeline includes steps of data collection, exploration, cleaning, preprocessing, training/evaluation, and presentation of results. Evaluation of models is performed with train and test data subsets by hold-out validation or k-fold cross-validation. Examples are given in R codes with explanation of parameters and hyperparameters, under- and over-fitting, bias and variance by the outcome of classification and regression models.
A Novel Chaotic Interior Search Algorithm for Global Optimization and Feature Selection
Published in Applied Artificial Intelligence, 2020
Sankalap Arora, Manik Sharma, Priyanka Anand
Data mining is a prominent research area that combines traditional data analysis techniques with emerging computational algorithms to assist in collecting heterogeneous data from distinct sources, transform it into valuable information and utilize it in designing effective business strategies for an enterprise (Han, Pei, and Kamber 2011). It has been widely used in several classifications, clustering, association and regression-based real-life problems of different domains. Feature selection is one of the challenging tasks of data quality which assists in selecting optimal data set’s features that boost the performance of classifiers. In other words, it is an extraction process which eradicates the irrelevant and redundant elements for the better understanding of data sets. The attributes that have a linear relationship with one another are called redundant attributes. The inclusion of these attributes seems to be unreasonable, as one can extract the complete information by merely incorporating one of these redundant attributes (Guyon and Elisseeff 2003). Nowadays, feature selection becomes mandatory as it is difficult to mine and transform the momentous volume of data into valuable insights.