Science topics: Data MiningClassification
Science topic
Classification - Science topic
The systematic arrangement of entities in any field into categories classes based on common characteristics such as properties, morphology, subject matter, etc.
Questions related to Classification
Selecting appropriate attributes is a vital step in the identification and classification process. In the field of non-destructive testing (NDT), this task is further complicated by limited sample availability and the lack of a representative database. In my paper titled “Evaluation of Reinforced Concrete Structures with Magnetic Method and ACO (Amplitude-Correlation-Offset) Decomposition”
I introduced the ACO decomposition technique, which enables effective identification via pattern matching (reference sample) and can also be used in classic classification. Additionally, I proposed a classification of feature-extraction methods in signals, highlighting the advantages and disadvantages of specific methods. What other strategies could be used to overcome the challenge posed by the absence of a statistically significant database?
Hello everyone,
I am currently trying to reproduce the results of a study on sleep stage classification using deep learning. The original paper reports an accuracy of around 80%, but despite following the described methodology, my model achieves significantly lower performance.
I would appreciate any guidance from researchers who have experience with sleep staging or EEG-based classification.
In particular, I am wondering about:
- important preprocessing steps (filtering, normalization, window size)
- recommended train/validation/test split for sleep datasets
- common pitfalls when working with datasets like Sleep-EDF or SHHS
- whether additional steps or hyperparameters are usually needed beyond what papers describe
Any insights or suggestions would be very helpful.
I am currently conducting research on integrating variational quantum algorithms with classical deep learning models to overcome the challenges of training high-dimensional networks. In my work, I am exploring whether quantum subroutines; such as quantum amplitude amplification or quantum natural gradient methods; can help speed up convergence and escape local minima more effectively than classical optimizers.
- Specific issues: I’m concerned with the effects of barren plateaus, noise accumulation, and limited coherence times on variational quantum circuits used for optimization.
- Research aspects: How do these hybrid approaches perform in terms of convergence rate and solution quality on realistic NISQ devices? Are there any demonstrated error mitigation techniques or circuit designs that help preserve gradient information in deep networks?
I would appreciate detailed theoretical analyses, simulation studies, or experimental benchmarks that compare these hybrid methods with traditional deep learning optimizers.
Sustainability reporting is an important aspect of corporate social responsibility (CSR), enabling companies to demonstrate their commitment to issues beyond financial performance and profit. The purpose of this study is to analyse business development sustainability reporting and justify its use as an effective CSR tool. During the study, the theoretical foundations of corporate reporting were further developed by clarifying the thesaurus on accounting and reporting. The refinement of the terminology framework was complemented by the justification of updated classification criteria and types of corporate reporting in the context of sustainability requirements. The classification criteria and types of corporate reporting were justified by the following: (1) the type of indicators (financial or non-financial), (2) the standard of non-financial reporting, (3) the areas of disclosure (management, environmental, social or financial), and (4) the type of non-financial disclosure (sustainability information as part of the annual report or a separate sustainability report).
what are the current challenges of using AI tools in medical image analysis such as burn skin injuries classification?
- Which deep learning architecture works best for burn image classification? Is deep learning can works with limited dataset?
What is the best classification AI application in order to achieve the accuracy result?
I want to classify pottery fragments, so I have color and texture features for 50 Images this means I have 50 vectors for classification. Which classification method is most effective for achieving the accuracy result?
National Heart, Lung, and blood Institute (NHLBI
ensemble classification will provide better performance than a single classifier. but the doubt is how o implement this one?
whether comparison will be done between two classification and the classification with best accuracy will be chosen or something else
I want to classify pottery fragments, so I have color and texture features for 50 Images this means I have 50 vectors for classification. Which classification method is most effective for achieving the accuracy result?
Hello colleagues,
I am currently working on a research project related to medical image analysis, specifically classification tasks using convolutional neural networks (CNNs). While the model achieves reasonable accuracy, I would like to further improve its performance.
I would appreciate your insights on the following:
- Which preprocessing techniques (e.g., normalization, augmentation, denoising) are most effective for medical images such as MRI or CT scans?
- Are there recommended CNN architectures (e.g., ResNet, DenseNet, EfficientNet) that tend to perform better in medical image classification compared to standard models?
- What regularization methods (dropout, weight decay, early stopping) or optimization strategies have you found helpful in avoiding overfitting with relatively small datasets?
- Would transfer learning from large natural image datasets (e.g., ImageNet) significantly improve performance in medical imaging tasks?
- Any advice on evaluation metrics that go beyond accuracy (such as F1-score, AUC, or sensitivity/specificity), which are particularly important in healthcare applications?
Thank you in advance for sharing your experience and recommendations!
In light of the recent uproar among recreational video game enthusiasts over the price increase of a popular title to $90 — a reaction I find justified due to repeated monetization of the same product — a more pressing ethical concern, in my view, lies elsewhere.
Austrian researcher Dr. Franz Schelling made a compelling ethical observation, stating:
"In view of the given facts we must ask ourselves whether the MS patient's shameless exploitation for an as senseless as profitable drug experimentation, grounded in an illogical MS 'definition' and 'identification', does not form an ethical issue."
I fully agree with Dr. Schelling’s position. The ongoing pharmaceutical experimentation on multiple sclerosis patients, often based on questionable theoretical assumptions and ambiguous disease classifications, seems to raise serious ethical concerns regarding human dignity and scientific integrity.
How do you evaluate this situation from a bioethical perspective? Do current research frameworks adequately safeguard the rights and integrity of MS patients, or is there a systemic failure that needs to be addressed?
I am trying to perform machine learning (classification) on a public health dataset for diagnosing a disease status (Positive, Negative). In my dataset, all of the features are categorical like educational level (Illiterate, Primary, Secondary, Higher), gender (Male, Female), Wealth index (Poor, Middle, Rich), Marital status (Unmarried, Married, Divorced, Separated, Widowed), etc. Many researchers before performing machine learning, they convert these features into numeric variables. Then, they standardize the features by transforming it into zero mean and unit variance. Since my background in Statistics, it seems quite illogical to me.
I have applied few tree based models (DT, RF, GBM, XGBM) and non-tree based models (SVM, KNN) using categorical variables without standardization, and got performance around 0.75 to 0.85 for accuracy, precision, sensitivity, specificity, and AUC.
Now, I have the following question:
- Does standardization of numerically encoded categorical feature logical, and hold any statistical validity??
I applied several machine learning classifiers (Decision Trees, Random Forest, GBM, XGBoost, SVM, and KNN) to a public health dataset for disease status classification. All features were categorical (e.g., education level, gender, wealth index). I did not encoded these categories as numerical values (e.g., Illiterate=1, Primary=2, etc.), and also did not standardize these categorical features.
I expect suggestions in this situation from the scientific community who are expert in this field. What should I do in this case??
I'm exploring optimal CNN architectures for breast cancer detection using deep learning. I want insights into effective feature selection techniques and hyperparameter tuning strategies to enhance model performance and generalization. What are the best approaches for fine-tuning pre-trained models for medical imaging?
Hello everyone! I am researching the applications of Vision-Language Models (VLM) in incremental learning for segmentation.
- While I've found many papers on unsupervised domain adaptation and image classification using VLM, there seems to be a lack of literature specifically on incremental learning segmentation. Could anyone recommend relevant papers or studies? Thank you very much!
What are the main challenges in training AI models for automatic classification and escalation of IT incidents in ITIL-based environments?
Given that the application of Artificial Intelligence (AI) for the automatic classification and escalation of incidents in ITIL-based environments presents various technical and operational challenges. These challenges directly impact the efficiency, accuracy, and reliability of AI models.
Deep Learning is better for image classification in most real-world scenarios, especially when:
· You have a lot of labeled image data
· You need end-to-end learning (raw pixels to predictions)
· Complex image patterns (e.g., object detection, medical imaging) need to be captured
Exceptions where traditional ML (like SVM) might be useful:
· Small datasets with high-quality hand-engineered features
· Simpler tasks (e.g., binary classification of simple images)
· Limited computing resources
The new classification been published by ILAE in 2025
Asai T. Taxonomic studies on acetic acid bacteria and allied oxidative bacteria isolated from fruits. A new classification of the oxidative bacteria. Journal of the Agricultural Chemical Society of Japan 1935; 11:674-708.
🧠 Feedforward Neural Networks (FNNs) Explained – In the Simplest Way!
🔗 https://youtu.be/C94d_NHtVSQ
Curious about how machines "learn"? 🤖
Start your neural network journey by understanding Feedforward Neural Networks (FNNs) — the foundation of deep learning!
In this video, you'll learn:
✅ What FNNs are
✅ How they process information
✅ Key components like layers, weights, and activations
✅ Their role in classification & prediction tasks
Perfect for beginners looking to build strong AI/ML fundamentals!
🎓 No prior deep learning experience required — just curiosity.
#MachineLearning #DeepLearning #NeuralNetworks #AI #Feedforward #FNN #ArtificialIntelligence #EdTech #LearningAI
Hello, my dear professors. Is there a researcher interested in wildlife and classification?
Details (provide context and specifics):
I’m working on a convolutional neural network (CNN) model for image classification using a small dataset (~1,000 samples). I’ve read that dropout can help prevent overfitting, but I’m unsure how to choose the right dropout rate.
- Are there best practices for setting the dropout rate in such scenarios?
- How sensitive is CNN performance to this hyperparameter when training data is limited?
- Any recommendations on validation approaches to tune dropout effectively?
I’d appreciate references or experiences from similar projects. Thank you!
Support Vector Machines (SVMs) and Convolutional Neural Networks (CNNs) are both used for classification tasks, including image classification. However, CNNs are generally better for image classification due to the following reasons:
· Automatic Feature Extraction
CNN automatically learns spatial hierarchies of features (edges, textures, shapes) from raw image pixels using convolutional layers. Whereas SVM Requires manual feature extraction(e.g., SIFT, HOG) before classification. The quality of results heavily depends on the features provided.
· Scalability with Large Datasets
CNN Scales well with large datasets and improves with more data.Whereas SVM Struggles with large datasets and becomes computationally expensive as the number of samples increases.
· Hierarchical Learning
CNNs can learn complex and deep feature representations, useful for distinguishing fine-grained patterns in images (e.g., faces, digits, animals). Whereas SVMs are shallow learners—they don't learn multi-level representations.
· Adaptability with Transfer Learning
CNNs can leverage pre-trained models like ResNet, VGG, etc., via transfer learning, significantly improving performance with fewer labeled images.Where as SVMs don’t benefit from this in the same way.
I am researching quantum kernel methods for classification tasks, focusing on mapping classical data into high-dimensional Hilbert spaces using parameterized quantum circuits. My goal is to design quantum feature maps that not only leverage quantum parallelism but are also robust against noise and decoherence inherent in current quantum hardware.
- Design challenges: What are the best practices for constructing quantum feature maps that maintain high fidelity and generalize well in the presence of NISQ-level noise?
- Performance evaluation: How can we quantitatively compare the classification performance of quantum kernels against classical kernels like RBF or polynomial kernels?
I am looking for theoretical frameworks, empirical studies, or benchmark experiments that address these challenges and offer guidelines for practical implementations in real-world high-dimensional datasets.
If a nickel mine is said to process ore from the 'smectite' and 'ferruginous' zones, does that include 'limonite' ore, or is limonite a completely different zone? What is the difference between these classifications? Thank you all
Dear RG members, IUGS is planning a new edition of the classical "Le Maitre" book devoted to the classification and nomenclature of igneous rocks. A group of 17 igneous petrologists (hereafter TGIR - Task Group on Igneous Rocks) is working for four years to update specific definitions or proposing entirely new sections.
As the Chair of the TGIR, I would like to start a discussion with all the interested people that want to give help concerning this task. Attached you find the first draft of Chapter 4 of the new book. I invite all the interested researchers to download it and add notes and comments using the "track changes" option with MS Word. I would appreciate then if you could send it to my email address (michele.lustrino@uniroma1.it), so that I can forward them to the TGIR members.
Thanks for the cooperation,
Michele
what is the best hierarchical image classification method for classifying the type of rice disease type at level-1 and further categorized into many categories based on the severity of the disease at level-2.
Dear researchers,
I’m working on a review paper about FRP bonding techniques for concrete strengthening, focusing on anti-debonding strategies. I’m struggling with the classification framework: existing studies mention mechanical anchorage, chemical modification, and hybrid connections, but the criteria vary
Questions:
1. Is there a widely accepted classification standard for FRP bonding techniques?
2. Should the review prioritize engineering application scenarios or mechanical principles?
Any recommended literature or advice would be greatly appreciated!
I have created LULC classification model. I want to know whether we can use total operating characteristic curve to validate the accuracy of LULC classification model?
Hi everyone,
We’re implementing the Track Quality Index (TUG_TQI) from Graz University of Technology to evaluate track conditions on Jakarta’s Light Rail Transit (LRT) network. The method aggregates track geometry parameters (gauge, cant, twist) into a single index, but we’re facing a couple of challenges:
1. Adapting TUG_TQI to Local Conditions
- Jakarta’s LRT has tight curves and low speeds (≤80 km/h).
- We collect data using a continuous measurement trolley, but longitudinal level (versine) is measured manually, leading to gaps.
Question: How can we tweak the TUG_TQI formula to work with fewer parameters and discontinuous versine data while keeping it reliable?
2. Establishing Track Quality Classifications
TUG_TQI itself doesn’t define quality thresholds like “good,” “fair,” or “poor.” We’re exploring ways to set these thresholds, such as:
- Statistical methods (e.g., quartiles, standard deviations of TQI distributions).
- Historical correlation, linking TQI to past maintenance records.
Questions:
- Are there case studies on defining TQI thresholds for similar networks?
- How can we adjust normalization and aggregation methods when data is incomplete?
- Would a hybrid approach (e.g., mixing EN standards with statistical analysis) be a good way to improve classification accuracy?
Any insights, references, or examples from similar rail systems would be greatly appreciated!
Thanks in advance for your thoughts.
References:[1] Offenbacher, S.; Neuhold, J.; Veit, P.; Landgraf, M. Analyzing Major Track Quality Indices and Introducing a Universally Applicable TQI. Appl. Sci. 2020, 10, 8490. https://doi.org/10.3390/app10238490
I have developed a model for LULC classification. How can I evaluate the performance of the model using TOC? I have calculated the confusion matrices for different thresholds for each class. I want to use the size of Misses, Hits, and False Alarms. As I have more than two categories, so my model validation will also have a component called Wrong Hits.
Looking for a deep learning layer, that is Keras compatible and could deliver high classification results from training and inferring on sequential data.
Besides having low inference latency and using the lowest computation possible.
Any suggestions are welcome including full models such as transformers and others.
Impotent: Stable Analysis Patterns for Software and Systems book will be awarded for the best answers.
M. E. Fayad. "Stable Analysis Patterns for Software and Systems" Boca Raton, FL: Auerbach Publications, Taylor & Francis Catalog #: K24627, May 2017. ISBN-13: 978-1-4987-0274-4
NOTES:
(1) There are many concept classifications from one field to another.
(2) Why is concept classification important?
Concepts are the building blocks of any field of knowledge.
(3) Every concept classification is based on different criteria from one field of knowledge to another. Consequently, each concept class has unique characteristics.
(4) Concept classifications create many significant problems: (a) misunderstanding between different Professions and (b) name concept instance, not the concept itself in most classifications.
Example: Exist concept classification is based on MODELING only
First
[I] Single Concept is a noun or concept in modeling a CLASS.
Such as Any Project, Any Proposal, Any Risk, and others
[II]. Compound Concept consists of 2 or more single concepts, called in modeling an ARCHITECTURAL CLASS, such as Data Analysis, Risk Assessment, and Machine Learning.
Second
[I] System Class is a noun or concept in modeling a CLASS – Specific Named Class, Such as sky city 1000, International Space Station, World War II,
[II] Actor/Party
Such as Person, Smart hardware, and Specific Creature type.
[III] Role
Such as Named Professor, Student, Dean,
Unified Concept Classification based on Fayad's Art of Abstraction
1) Enduring Business Theme (EBT) -- Rank -- highest Concept
2) Business Object (BO) -- Rank -- Very Important
3) Industrial Object (IO) -- Rank -- Changeable
+ Strong and complete
+ Unified -- 1 and 2
+ Stable -- 1 and 2
+ All the concepts above in our classification are single concepts.
+ Architectural Concepts are a combination of
-- Two or more EBTs
-- Two or more BOs
-- Mix of two or more EBTs + Bos
Notes:
(1) General purpose concept classification
(2) Unified
(3) Stable Overtime
(4) Base of understanding by different Professions
(5) Dealing with the actual concepts
(6) No Synonymous
Where:
Enduring Business Themes (EBTs)
1. Unified
2. Stable
3. Final
4. Continuous
5. Ultimate Goals
6. Has rules of conduct -- (High level) -- Rules must be known to all. -- Unfortunately, many people do not know them.
7. Important discoveries
8. Examples -- Friendship, Marriage, Thinking, Retaliation.
Business Objects (BOs)
1. Unified
2. Stable internally and adaptable externally
3. Final
4. Has a beginning and an end
5. Has an ultimate goal that can be positive or negative
6. We add the word "Any" to each of these concepts
7. Has rules of conduct -- (moderate level) -- Rules must be known to all. -- Unfortunately, many people do not know them.
8. Important discoveries
9. Examples -- Any Project, Any Proposal, Any Data.
Industrial Objects (IOs)
1. Changeable
2. Tangible
3. Unfortunately, and currently building and developing everything based on them. (Disasters)
3. Application Objects
4. Well-known to the majority of people
5. Has no value -- The strange thing is people say, "I love my car or smoking."
6. Some have side effects or high impacts on society, such as Oil, Drugs.
7. Examples: Specific Novel, Conference Table, MacBook.
Please find the type (EBT, BO, IO) of each of the concepts following: (10 out of 13)
Culture
Privacy
Policy
Analysis
Brain
Guru
Intelligence
Assumption
World War II
Poverty
iPhone
Feedback
Bureaucracy
I am actively researching quantum reinforcement learning, aiming to leverage quantum-enhanced decision-making for sequential tasks. Although quantum algorithms promise speedups through techniques like quantum random walks or Grover’s search for exploration, practical QRL implementations on NISQ devices face significant hurdles due to noise, decoherence, and hardware limitations.
- Technical challenges: How do noise and limited qubit connectivity affect the convergence and stability of quantum reinforcement learning algorithms?
- Mitigation strategies: What error mitigation techniques; such as zero-noise extrapolation, dynamical decoupling, or hybrid feedback loops; are most effective in preserving quantum advantages during the iterative learning process?
- Implementation insights: Additionally, I seek guidance on designing robust quantum reward functions and state encoding methods that can adapt to hardware imperfections.
Any detailed case studies, simulation results, or experimental benchmarks from platforms like IBM Q, Rigetti, or photonic quantum processors would be extremely valuable for advancing this research.
How to prepare a Shukalev classification chart/table to define the groundwater types?
estuarine sediment classification diagram is known as sheperd diagram. And it classifies sediement into sand silt and clay. Is there anyone who knows how to draw that diagram ?
Dear RG members, IUGS is planning a new edition of the classical "Le Maitre" book devoted to the classification and nomenclature of igneous rocks. A group of 17 igneous petrologists (hereafter TGIR - Task Group on Igneous Rocks) is working for four years to update specific definitions or proposing entirely new sections.
As the Chair of the TGIR, I would like to start a discussion with all the interested people that want to give help concerning this task. Attached you find the first draft of Chapter 3 of the new book. I invite all the interested researchers to download it and add notes and comments using the "track changes" option with MS Word. I would appreciate then if you could send it to my email address (michele.lustrino@uniroma1.it), so that I can forward them to the TGIR members.
Thanks for the cooperation,
Michele
I am studying two groups of patients group A that did not need surgery, and Group B that required surgery. the outcome variable is a classification which has four ordinal classes (One through Four with worsening outcomes as grades increase).
Which statistical test will fit the bill the most?
Dear, I have a question: why doesn’t this link work when I search for it on Google?
Enhancement of Breast Cancer Classification Using Bat Feature Selection with Recurrent Deep Learning
Why doesn’t this address the paperwork when I search for it on Google? Enhancement of breast cancer classification using bat feature selection with recurrent deep learning.
I think the reason is in the Google's search engine and your region or/and your VPN (if any).
From my place I can see the working links (see attachment).
Hi, I'm currently doing a research for my thesis, can someone tell me where can I find publicly available datasets for baby crying classification (e.g., SPLANN, Baby2020, Baby Chillanto, DSPLabCup)? I’ve tried contacting the authors of these datasets without success. If anyone has access to these or similar datasets or knows where I can download them, I’d greatly appreciate your help!
Thanks,
Have a nice day!
Dear RG members, IUGS is planning a new edition of the classical "Le Maitre" book devoted to the classification and nomenclature of igneous rocks. A group of 17 igneous petrologists (hereafter TGIR - Task Group on Igneous Rocks) is working for four years to update specific definitions or proposing entirely new sections.
As the Chair of the TGIR, I would like to start a discussion with all the interested people that want to give help concerning this task. Attached you find the first draft of Chapter 2 of the new book. I invite all the interested researchers to download it and add notes and comments using the "track changes" option with MS Word. I would appreciate then if you could send it to my email address (michele.lustrino@uniroma1.it), so that I can forward them to the TGIR members.
Thanks for the cooperation,
Michele
Dear RG members, IUGS is planning a new edition of the classical "Le Maitre" book devoted to the classification and nomenclature of igneous rocks. A group of 17 igneous petrologists (hereafter TGIR - Task Group on Igneous Rocks) is working for four years to update specific definitions or proposing entirely new sections.
As the Chair of the TGIR, I would like to start a discussion with all the interested people that want to give help concerning this task. Attached you find the first draft of Chapter 5 of the new book. I invite all the interested researchers to download it and add notes and comments using the "track changes" option with MS Word. I would appreciate then if you could send it to my email address (michele.lustrino@uniroma1.it), so that I can forward them to the TGIR members.
Thanks for the cooperation,
Michele
Dear RG members, IUGS is planning a new edition of the classical "Le Maitre" book devoted to the classification and nomenclature of igneous rocks. A group of 17 igneous petrologists (hereafter TGIR - Task Group on Igneous Rocks) is working for four years to update specific definitions or proposing entirely new sections.
As the Chair of the TGIR, I would like to start a discussion with all the interested people that want to give help concerning this task. Attached you find the first draft of Chapter 6 of the new book. I invite all the interested researchers to download it and add notes and comments using the "track changes" option with MS Word. I would appreciate then if you could send it to my email address (michele.lustrino@uniroma1.it), so that I can forward them to the TGIR members.
Thanks for the cooperation,
Michele
Classification or clustering algorithms and how to apply them to existing big data from social networks can help in this. Precise methods and work processes are desired.
Does classifying the researcher according to his beliefs affect the extent to which the results of his research are accepted?
I am exploring the CutMix and MixUp data augmentation methods in the context of computer vision tasks. Could you explain the key differences between the two techniques, and in which scenarios one method might be more effective? Additionally, I am interested in understanding the different use cases for each approach in improving model generalization and performance, especially in tasks like image classification, object detection, and segmentation. Any insights or recent research on their applications would be greatly appreciated!
Diversity of classifications for both leadership and management according to the characteristics of the manager or leader. Are these classifications considered a diagnosis of a person? How can a manager or leader be described if he does not have a fixed pattern, especially since changing his characteristics is not necessarily proportional to the situations the organization is going through?
In a broad sense, can all the deep learning tasks be viewed as classification?
I need the full text of several books/journal articles listed below, as these references are essential for my taxonomic study of Lycopodiaceae in the Indonesian region. The required works are as follows:
- Herter, G. 1949a. Index Lycopodiorum. Montevideo, Uruguay: Herbarium Herter.
- Herter, G. 1949b. Systema Lycopodiorum. Revista Sudamericana de Botanica 8: 67-86.
- Herter, G. 1950. Systema Lycopodiorum. Revista Sudamericana de Botanica 8: 93-116.
- Øllgaard, B. 1987. A revised classification of the Lycopodiaceae s.lat. Opera Botanica 92: 153-178. ISSN: 0078-5237
- Wilce, J.H. 1961. Lycopodium complanatum L. and four new allied species of the Old World. Nova Hedwigia 3: 93-117. ISSN: 0029-5035
- Wilce, J.H. 1965. Section Complanata of the genus Lycopodium. Beih. Nova Hedwigia 19: 1-233. ISSN: 0029-5035
If you have any information or access to these materials, it would greatly help me, as all these references are inaccessible from my country. Thanks.
Dear Colleagues,
I am contributing to a book titled "Microbial Enzymes: Classification, Biochemistry, Production, and Applications," edited by Prof. Dr. Abdel Moneim Elhadi Sulieman.
I am currently working on the following two chapters:
- Mechanism of Microbial Enzymes
- Bacterial Enzymes and Their Applications
I am seeking two collaborators (one for each chapter) who are interested in co-authoring with me. The deadline for submission is January 15, 2025. I will accept the first two qualified respondents.
If you are interested, please let me know as soon as possible.
Best regards,
Prof. Dr. Emad Abdallah
Microbiologist
particularly models that achieve similar performance levels.
Hello Everyone
I have a classification dataset and I want to extract a Boolean Function for it using Genetic Programming. first of all is it possible?
could you please advise about how can I do that. I'm new to this field (GP). Therefore, any resources will be helpful.
please also note that, features in this dataset are discrete numbers.
I want to gather data on how many research papers have been published from 2017 to 2024 on the topic "breast cancer classification on pathological images using deep learning models." How can I do this?
I want to know more about classification of igneous rocks.
I think ASA classification needs to be revised and updated because there are many important factors like patient's age and difficult airway that affects so much and increse the risk of the Anesthesia and operation risk on the patients are not included or considered in the classification .
🚀 Unlock the Power of Competitive Neural Networks! 🚀
Are you ready to dive deep into the fascinating world of Neural Networks? 🤖
Check out my latest video https://www.youtube.com/watch?v=lD0yLzMPJbM where I explain the concept of Competitive Neural Networks, and how they stand apart in the realm of Machine Learning! Learn how these networks specialize in clustering and classification tasks. 🧠✨
🔗 Watch now: Competitive Neural Networks
Don’t forget to subscribe for more AI, ML, and cutting-edge tech insights! 💻🔍
#MachineLearning #NeuralNetworks #AI #ArtificialIntelligence #DataScience #DeepLearning #CompetitiveLearning #ProfessorRahulJain #TechEducation #AIResearch
Dear RG members, IUGS is planning a new edition of the classical "Le Maitre" book devoted to the classification and nomenclature of igneous rocks. A group of 17 igneous petrologists (hereafter TGIR - Task Group on Igneous Rocks) is working for three years to update specific definitions or proposing entirely new sections.
As the Chair of the TGIR, I would like to start a discussion with all the interested people that want to give help concerning this task. I and the other members of the TGIR will start posting a series of arguments that will greatly benefit from your comments, so I hope to receive stimulating feedback.
The IUGS classification does not include eclogites among igneous rocks.
Eclogite (Haüy, 1892): Rock composed by grass-green pyroxene (omphacite) and reddish/purplish garnet.
Eclogite facies (Eskola, 1921): Plagioclase-free high-pressure and high-temperature rocks, with mafic protolith (often with basaltic composition), with mineralogy represented mainly by omphacite (Na-Ca-Mg-Al-rich clinopyroxene) + pyrope (Mg-Al-rich garnet), usually with granoblastic texture.
Eclogite (IUGS Desmond and Fettes, 2007): Plagioclase-free metamorphic rock composed of ≥75 percent vol. of omphacite and garnet, both of which are present as major constituents, the amount of neither of them being higher than 75 percent vol.
Most of the basaltic melts do not reach Earth's surface. Those solidifying at >1-1.5 GPa crystallize out of the plagioclase stability limit. Paradoxically, a basaltic melt will crystallize in a plagioclase-free mineral assemblage. In any case, being this rock with basalti composition associated to solidification of a magma, it should be classified among igneous rocks. Also chemically it should be classified as baasalt, because plotting in the TAS basalt field. However, from a petrographic point of view, it cannot be classified as basalt (because ultramafic, i.e., with the sum of Q-A-P-F minerals <10% and, above all, being plagioclase-free). Petrographically it should be classified as eclogite (garnet + Na-rich cpx), but at the same time it should be classified also as a basalt.
We propose to add a comment that eclogitic rocks should also have an igneous origin
Comments welcome,
Michele
I think that ASA classification needs to be revised and updated to include and consider other important factors like patient age especially the extremes( neonates and old aged people) , and also other important factor difficult airway,
I am trying to solve the Wi-Fi offloading decision making problem using classification and clustering of known and unknown traffic respectively in a given mobile network using bi-flows of packets in the network
I am considering taking as input A and A' with some modification to A, and classifying the classes based on the relationship between the two. Does such a method exist?
(A method to perform classification based on the relationship between the two inputs)
Do you know of a database that has classified chemical compounds (classified like plants)? Unfortunately, the websites introduced in the article "ClassyFire: automated chemical classification with a comprehensive, computable Taxonomy" they are out of reach
Is it correct to use metrics such as the number of articles, citations, H index, and I10 index to classify researchers? Don't researchers compromise on quality and sometimes even ethical values to increase these metrics? Don't you think these measurements are unscientific?
Hi all,
I am attempting supervised and object-based image classification using Arch Pro (V.3.2). Here are the steps I have followed:
I acquired Sentinel-2 imagery and combined bands 2, 3, 4, and 8 (10m resolution).
I performed image segmentation.
I created a classification schema.
I generated training samples.
However, when I attempt to classify using a support vector machine classifier, I encounter the following error:
ERROR 003436: No training samples found for these classes: Soil, Water, Impervious, Grass, Tress.
The table was not found. [VAT_Segmented_202407110934456475089_interIndex]
The table was not found. [VAT_Segmented_202407110934456475089_interIndex]
Failed to execute (TrainSupportVectorMachineClassifier).
Failed at Thursday, July 11, 2024 9:44:52 AM (Elapsed Time: 0.22 seconds)
(I have attached a screenshot of the error.)
I have tried several times but haven't been able to identify the cause of this error. Do any of you know what might be causing it?
To cite my article, I need some reference papers based on only CNN models using 5 or 7 convolution layers for the breast cancer binary classification in the BreakHis histopathology image dataset.
Dear colleagues,
Is it possible to send me the list of journals impact factor for the year 2024 (classification is for the year 2023)? excel format if it is possible.
Thank you in advance.
Sincerely,
Yassine.
What are the new and recent models are available for breast cancer binary classification
Greetings! I would like to find sources for a project, related to the classification and segmentation of sports betting players. Thank you very much for the suggestions!
How can attention mechanisms be integrated with convolutional neural networks to enhance performance in image classification tasks?
How does the application of generative adversarial networks (GANs) for data augmentation impact the robustness and accuracy of image classification models?
I'm started a discussion on the topic of "Which methods can be implemented for the classification of cities in different countries of the world?"
Hi everyone, I'm working on skin cancer classification and I've extracted the feature from three pre-trained CNN models and concatenated all the features. Finally a dense layer with softmax function was used for classification.
But I encountered constant validation accuracy with lower training accuracy.
How can I solve this problem please. Despite this, I have tried many optimizer functions with different learning rate, regularizer, early_stopping, dropout, and also I used image augmentation
Hi folks,
I'm a computer scientist PhD student, and I'm working on implementing Multi-Task Learning architecture for a better generalization aims, it will be throughout a Deep Learning model.
I have some questions concerning MTL algorithms and its feasibility, for those whom already worked on the same project, here are my questions:
1- Can we design an MTL architecture model based on different task's definition ? Example: task 01: is a classification, task 02: is a clustering (mixing between supervised and unsupervised tasks) is it possible, or we have to design a common and homogeneous architecture ?
2- Is it a mandatory to assign a specific dataset for each task ? Or, we can use a common and global dataset for both shared layers and tasks specific layers (example: an ecommerce historical purchase) ?
3- According to you, what are the best pretraining MTL architecture models that I could rely on ?
Thanks in advance !
Hello everyone,
I have a dataset of videos for action classification, where each video contains multiple actions. I need to annotate these videos with the name of each class and the start and end times of each action. Does anyone know the fastest and most straightforward way to manually label this?
Thank you!
What is scope of the implementing LIS classification and cataloguing in different field ?
Keeping Murphy's Law, the KISS principle, and Popper's Logic, right, left, and center, respectively, as well as the all-encompassing muses of insight, innovation, intuition, imagination, and insurrection (the 5 I's that holistically, through immersive-integrative multi-disciplinary contemplative approach identifies the noise or separates the wheat-from-the-chaff at the intersection of fact and fiction), and importantly and synergistically compose the whole that is greater than the sum of its parts -- in true Aristotelian fashion -- that govern progress and advance in human thinking through the synapse in all human endeavor, scientific and non-scientific.
I will put exactly 50-years of my part in one of the greatest mysteries ever faced by humans, and that will follow this species indefinitely to perpetuity, but with secure and fearless knowledge through application of principles or laws of theory and therapy, elimination of canonical or Institutional myths and assumptions, with a complete unwinding of this humungous Gordian knot of neuro-ophthalmology.
Da Vinci guarded against excessive use of words to describe any entity or anything. Migraine is an entity of excess -- incidence, words, data, statistics, analyses, meta-analyses, hypotheses, viewpoints, perspectives, Editorials, Medical Conferencing Abstracts, invited Lectures, hyper-splitting of nosology, and Letters-to-the-Editor, all claiming to know a slice of truth or presumed truth about migraine with a hyper-exponential absolutely unlimited untrammeled expansion. Quo vadis is not even a remote concept.
I, in the Third Millennium, describe the 'what' of migraine in 6-10 words, a definition that will last to perpetuity:
Migraine is the delayed outcome of an oculo-cephalic autonomic storm (causing the non-homonymous scintillating scotoma as well as the lateralizing headache). More succinctly, migraine is an oculo-cephalic autonomic storm.
Nothing is static. No theory or therapy cannot be improved. The core of migraine is here.
With the cause-effect mechanisms in migraine pathophysiology fully described, what has been missing for 6 millennia or more is presented right here and now.
The doors of perception for cluster headache and other indomethacin-responsive headaches are now open.
Reversal of the hyper-split classification of primary headaches is imminent, leading to a holistic comprehensive understand of a large section of medicine and neuroscience.
06-MAY-2024
New Delhi
ORCID iD: 0000-0002-6770-5916
Within a project about geographical traceability of horticultural products, we would like to apply classification models to our data set (e.g. LDA) to predict if it is possible to correctly classify samples according to their origin and based on the results of 20-25 different chemical variables.
We identified 5 cultivation areas and selected 41 orchards (experimental units) in total. In each orchard, 10 samples were collected (each sample from a different tree). The samples were analyzed separately. So, at the end, we have the results for 410 samples.
The question is: the 10 samples per orchard have to be considered pseudoreplicates since they belong to the same experimental unit (even if collected from indepedent trees)? Should the LDA be performed considering 41 replicates (the 41 orchards, taking the average of the 10 samples) or should we run it for the whole dataset?
Thank you for your help.
Please provide an explanation according to the classification of primary and secondary uranium ore
Dear researchers,
a variability in the taxonomy classification of microbial communities when using different primer pairs (e.g. for 16S rDNA) is commonly known. However, the mismatches to these primers are not described as the major reason for this bias. My question is: what are other possible causes of this bias and which is now supposed to be the major one?
Thank you for your contribution. Lucie
During my research for my bachelors thesis into the classification of mining in the Reichsgruppe under the Nazi regime, I came across two different classifications:
- On the one hand, Boris Gehlen in his chapter "3.10 Energy industry" (in https://www.degruyter.com/document/doi/10.1515/9783110796353-016/html?lang=de) assigned mining to the Reichsgruppe "Energy Industry", which sounds logical to me.
- On the other hand, I have found some works that explain that mining belonged to the Reichsgruppe "Industry": https://www.degruyter.com/document/doi/10.1524/jbwg.1980.21.1.177/html
So which is correct?
I can only explain this apparent double classification by the central importance of mining for both economic sectors: On the one hand as a supplier of raw materials for industrial production and on the other as a key sector for energy supply.
Is it possible that in different sources and at different times the emphasis was placed more on one or the other affiliation, depending on which aspect of coal mining was in the foreground, so that it can be said that mining belonged to both groups?
Should I use the traditional UHI index classification (<0, 0-0.005, 0.005-0.010, 0.010-0.015, 0.015-0.020) where 99% of my area falls under UHI effect, or would a natural classification or classification like this (<0, 0-0.05, 0.05-0.10, 0.10-0.15, 0.15-0.20, >0.20) be more suitable for studying the urban heat island effect in our tropical region?
ما هي معايير التصنيف الأكاديمي للجامعات؟
These beautiful specimens of Cerambyx were photographed by me, (not captured)
in Abruzzo (Central Italy) loc. Sulmona. I was able to identify the genus but not the species.
Entomologists help me, I am looking for an exact classification.
Supervised Learning
In supervised learning, the dataset is labeled, meaning each input has an associated output or target variable. For instance, if you're working on a classification problem to predict whether an email is spam or not, each email in the dataset would be labeled as either spam or not spam. Algorithms in supervised learning are trained using this labeled data. They learn the relationship between the input variables and the output by being guided or supervised by this known information. The ultimate goal is to develop a model that can accurately map inputs to outputs by learning from the labeled dataset. Common tasks include classification, regression, and ranking.
Unsupervised Learning
Unsupervised learning deals with unlabeled data, where the information does not have corresponding output labels. There's no specific target variable for the algorithm to predict. Algorithms in unsupervised learning aim to find patterns, structures, or relationships within the data without explicit guidance. For instance, clustering algorithms group similar data points together based on some similarity or distance measure. The primary goal is to explore and extract insights from the data, uncover hidden structures, detect anomalies, or reduce the dimensionality of the dataset without any predefined outcomes. Supervised learning uses labeled data with known outcomes to train models for prediction or classification tasks, while unsupervised learning works with unlabeled data to explore and discover inherent patterns or structures within the dataset without explicit guidance on the expected output. Both have distinct applications and are used in different scenarios based on the nature of the dataset and the desired outcomes.
I am looking for a high-resolution dataset (alternative to ImageNet) that has classes with sub-groups. I need this dataset for the domain transfer experiment. Basically, I will be using DNN pre-trained on ImageNet to extract features. Example: in CIFAR100, 100 classes are grouped into 20 super classes, such that each super classes have some sub-classes. I need similar dataset, but it has to be high resolution. Can you suggest any suggest one?
I am using Google Earth Engine for LULC classification map. For this purpose I have used smile random forest classifier to classify the Landsat 7 Top of Atmosphere data. Now could you please tell me how can I validate the LULC classification map?
What is the pixel classification for different land use in an NDBI map for hilly areas?
Can i cluster documents to label them as a first step. Then in the second step, can I use the labelled documents to apply a classification method such as svm, knn, etc.?
What is the most appropriate classification for the 12 identified compounds from the essential oil of a medicinal plant, which include:
1. Camphene,
2.Para-Cymene,
3. 1-Limonene,
4.Gamma-Terpinen,
5.Trans-Decalone,
6. Cuminic Aldehyde,
7. Cyclopentanone,
8. Acetyl phenyl carbinol,
9. 1-amino-1-o,
10. 5-Methyl-2-Phenyl indolizine,
11. Silicic acid,
12. 5-nitrobenzofuran-2-carboxylic
Is the categorization in the attached image correct?
Need help with an unsupervised deep image stacking project. Image stacking is a commonly used technique in astrophotography and other areas to improve the signal-to-noise ratio of images. The process works by first aligning a large number of short exposure images and then averaging them which reduces the noise variance of individual pixels. I have to do this process with neural networks by predicting a distortion field for each image and using a consistency objective that tries to maximize the coherence between the undistorted images in the stack and the final output. I need some learning materials for performing image stacking preferably in python and make a neural network. I already have experiences with training object classification and detection models and have worked on different YOLO models.
Hyperspectral Imaging, Hyperspectral Classification, Statistical Test
Hello everyone,
I have applied 1D CNN on speech emotion recognition, when I shuffled columns I got diffrent results, for example, using matrix(:,[1 2 3]) gives different classification results than matrix(:,[2 3 1]) which should be the same, I have tried rng("default") but it hasn't resolved the issue. Can someone please assist me with this?
Thank you in advance!
