Stock management and meat yield prediction in crayfish populations

Description

[0001] STOCK MANAGEMENT AND MEAT YIELD PREDICTION IN CRAYFISH POPULATIONS

[0002] Technical Field

[0003] The present invention relates to a system configured to perform stock management by predicting a population size, a growth performance, and a meat yield efficiency of crayfish populations by means of a machine learning algorithm, and further to enable sustainable fisheries, population, and aquatic resources management.

[0004] Background of the Invention

[0005] Freshwater lobsters, also known as crayfish, constitute one of the largest inland water forms of crustaceans within the order Decapoda, include economically significant species, and are represented worldwide by 737 species and subspecies. Globally, crayfish production is carried out through capture fisheries and aquaculture. Despite the large number of existing species, harvesting and production activities are generally concentrated on a limited number of economically significant families, namely Cambaridae, Parastacidae, and Astacidae. In the existing state of the art, stock estimation and meat yield prediction in crayfish are performed using conventional fisheries methods, classical statistical techniques, and individual experience.

[0006] In recent years, crayfish plague, overfishing, global warming, and increasing water pollution have led to considerable fluctuations in crayfish production. As a result of these fluctuating trends in crayfish production, stock management practices and alternative production methods must be improved and further developed, and available data must be used in a careful and effective manner. Accordingly, crayfish producers are required to provide data suitable for processing with modern technologies in order to support decision-makers in defining accurate future strategies and decisions. However, the manual processing and analysis of very large volumes of data is not feasible. Machine learning involves a wide range of approaches that incorporate patterns for establishing relationships, which in turn determine a technique to interpret an output derived from data. In general, the most commonly used machine learning methods include artificial neural networks, logisticregression, fuzzy modeling, genetic algorithms and genetic programming, decision trees, Bayesian network approaches, random forests, and support vector machines.

[0007] According to the known state of the art, regression-based modeling techniques are widely used in aquatic resources studies to predict species distribution and water quality. Such techniques include generalized additive models (GAMs), generalized linear models (GLMs), classification and regression trees (CART), and multivariate adaptive regression splines (MARS). In addition, a modeling approach that combines a functional network framework with a dynamic Bayesian network model is employed to predict trends in different fish and zooplankton species under specific fisheries, temperature, and net primary production (Net PP) scenarios.

[0008] Furthermore, Bayesian analysis is applied to develop habitat suitability models and to calculate the combined uncertainty and variability of parameters in crayfish bioaccumulation models.

[0009] In crayfish populations, serious challenges arise as a result of disease, fishing pressure, global warming, and environmental pollution as described above, and scenarios involving local species extinction may occur. Due to the advantages offered by machine learning applications, including the ability to evaluate complex and typically high-dimensional datasets that exhibit nonlinear dependencies among multiple variables and unknown interactions, as well as data that do not conform to the assumptions of conventional statistical methods, there is a need for a new evaluation system that contributes to sustainable and economically viable stock management and to the accurate prediction of meat yield quantities.

[0010] Description of the Invention

[0011] In the present description, the invention is provided solely for facilitating a clearer understanding of the subject matter, without any intention to impose a limiting effect.

[0012] The characteristics and structural features of the invention, along with all the advantages thereof, will be more clearly understood from the detailed description disclosed below; accordingly, the assessment should be made with due consideration of the detailed description.The present invention relates to a system for stock management and meat yield prediction in crayfish populations, which satisfies the aforementioned requirements, eliminates existing disadvantages, and provides additional advantages.

[0013] A primary object of the invention is to perform stock management by predicting a population size and a meat yield efficiency of crayfish populations by means of a machine learning algorithm, and further to enable sustainable fisheries, population, and aquatic resources management.

[0014] Another object of the invention is to determine a status of the population by maximizing the utilization of population data while minimizing costs.

[0015] Yet another object of the invention is to minimize the time and labor burden associated with population studies that are typically conducted over extended periods.

[0016] Still another object of the invention is to predict stock management and meat yield in crayfish populations and to support sustainable management thereof.

[0017] Preferably, the invention provides a system configured to perform stock management by predicting a population size, a growth performance, and a meat yield efficiency of crayfish populations by means of a machine learning algorithm, and further to enable sustainable fisheries, population, and aquatic resources management.

[0018] More preferably, the invention provides a system for determining and managing population status through the comprehensive utilization of population data.

[0019] More preferably, the invention provides a management system for stock management and meat yield prediction in crayfish populations.

[0020] A flow diagram of the invention illustrated in Figure 1 consists of two modules. In a first module (1), data preprocessing techniques are applied across all datasets. In a second module (2), data obtained from a dataset formed with a ratio of 70 percent training data and 30 percent test data are analyzed using Support Vector Regression (SVR), following whichthe machine learning model is optimized. Upon completion of the optimization process, meat yield and status of the population are predicted.

[0021] The abbreviations and corresponding definitions of the features used for stock management and meat yield prediction in crayfish populations are set forth below, in which:

[0022] SN: Sample number

[0023] CD: Capture day

[0024] G: Gender

[0025] CL: Cephalothorax length

[0026] CW: Cephalothorax width

[0027] AL: Abdominal length

[0028] AW: Abdominal width

[0029] CFL-R: Right chela length

[0030] CFL-L: Left chela length

[0031] SL-R: Right chela finger length

[0032] SL-L: Left chela finger length

[0033] SW-R: Right chela width

[0034] SW-L: Left chela width

[0035] CH: Cephalothorax weight

[0036] AH: Abdominal weight

[0037] AMH: Abdominal meat weight

[0038] SH-R: Right chela weight

[0039] SH-L: Left chela weight

[0040] SMH-R: Right chela meat weight

[0041] SMH-L: Left chela meat weight

[0042] TW: Total weight

[0043] TL: Total length

[0044] Following the input of feature data into the machine learning algorithm in accordance with the flow diagram illustrated in Figure 1 , analysis is completed by means of Support Vector Regression (SVR), so that meat yield efficiency prediction and stock management of the population is performed.Ultimately, the invention allows for the determination and management of reproductive activities, meat yield efficiencies, and population status of crayfish populations over time. Furthermore, time and labor losses arising from the prolonged duration of population studies required to analyze this status are eliminated. In addition, the invention enables sustainable population and aquatic resources management.

Claims

CLAIMS1. A system configured to perform stock management by predicting a population size, a growth performance, and a meat yield efficiency of crayfish populations by means of a machine learning algorithm, and further to enable sustainable fisheries, population, and aquatic resources management, characterized in that the system predicts the population size, population status, and meat yield efficiency of the crayfish population by means of the machine learning algorithm, while simultaneously enabling sustainable fisheries, population, and aquatic resources management.

2. A system according to claim 1, characterized in that the system is capable of analyzing data obtained from morphometric measurements of the crayfish.

3. A system according to claim 1 or claim 2, characterized in that the system uses, as a feature, data including SN (sample number), CD (capture day), G (gender), CL (cephalothorax length), CW (cephalothorax width), AL (abdominal length), AW (abdominal width), CFL-R (right chela length), CFL-L (left chela length), SL-R (right chela finger length), SL-L (left chela finger length), SW-R (right chela width), SW-L (left chela width), CH (cephalothorax weight), AH (abdominal weight), AMH (abdominal meat weight), SH-R (right chela weight), SH-L (left chela weight), SMH- R (right chela meat weight), SMH-L (left chela meat weight), TW (total weight), and TL (total length).

4. A system according to claim 1 or claim 2 or claim 3, characterized in that analysis is completed by means of Support Vector Regression (SVR) to provide meat yield efficiency prediction of the population and to present stock management results accordingly.

Images