Data Management ,Data Collection,Data Pre-processing for AI&ML in Healthcare

Data Collection,
Management, and
Pre-processing for
AI&ML in
Healthcare
By,
Prof.Jagruti Jadhav

• Importance of Data Analysis in Modern Healthcare
• Data analysis has become an indispensable tool in modern
healthcare, revolutionizing the way medical professionals diagnose,
treat, and manage patients. The sheer volume of healthcare data
generated daily, combined with advances in technology, has
highlighted the significance of harnessing data-driven insights to
enhance patient outcomes, streamline operations, and drive
medical advancements.

Data analysis is crucial in today's healthcare landscape as follows:
• 1. Informed Decision-Making: Data analysis empowers healthcare providers,
researchers, and administrators with evidence-based insights. By examining
patterns and trends within large datasets, decision-makers can make informed
choices about patient care, resource allocation, and strategic planning.
• 2. Personalized Medicine: Healthcare is moving towards personalized
treatment plans tailored to individual patient characteristics. Data analysis
enables the identification of patient subgroups with similar profiles, ensuring
treatments are more precise, effective, and well-suited to each patient's unique
needs.

• 3. Early Disease Detection: Data analysis can identify subtle patterns
indicative of early disease onset. By detecting anomalies or unusual
trends, healthcare professionals can intervene before conditions
worsen, leading to earlier interventions and improved patient
outcomes.
• 4. Predictive Analytics: Predictive models built through data analysis
can forecast patient outcomes, disease progression, and potential
complications. This information allows medical teams to take
proactive measures, reducing hospital readmissions and improving
patient care.

• 5. Efficient Resource Management: Data analysis optimizes
healthcare resource allocation. Hospitals can identify peak patient
demand times, allocate staff and resources accordingly, and
streamline processes for better patient flow.
• 6. Clinical Research and Drug Development: Data analysis
accelerates clinical research and drug development processes.
Researchers can analyze vast datasets to identify potential targets
for new drugs, streamline clinical trials, and predict drug
interactions and adverse effects.

• 7. Population Health Management: By analyzing demographic
and health data of populations, public health officials can design
targeted interventions, track disease outbreaks, and implement
preventive measures to improve community well-being.
• 8. Quality Improvement: Data analysis identifies areas for quality
improvement. By analyzing patient outcomes, medical errors, and
patient feedback, healthcare institutions can implement measures
to enhance the quality and safety of care.

• 9. Healthcare Cost Management: Efficient data analysis helps in
cost containment and waste reduction. It identifies unnecessary
procedures, overutilization of resources, and billing discrepancies,
leading to more cost-effective healthcare delivery.
• 10. Continuous Learning and Research: Data analysis facilitates a
continuous learning cycle. By analyzing clinical outcomes,
researchers can identify best practices, refine treatment protocols,
and contribute to the ongoing evolution of medical knowledge.

• Electronic Health Records (EHRs)
• Definition of Electronic Health Records (EHRs).
• Electronic Health Records (EHRs) refer to digital versions of patients'
medical history, health information, and other healthcare-related data
that are stored and managed electronically. EHRs provide a
comprehensive and centralized repository of a patient's medical and
clinical information, accessible to authorized healthcare professionals
across various healthcare settings. These records replace traditional
paper-based medical records, offering numerous advantages in terms
of efficiency, accuracy, and patient care.

• EHRs typically include a wide range of information, such as:
1.Patient Demographics: Basic patient information like name, date of birth,
contact details, and emergency contacts.
2.Medical History: Past medical conditions, surgeries, allergies, immunizations, and
family medical history.
3.Medications: List of current and past medications, dosages, and instructions.
4.Laboratory and Test Results: Results of blood tests, imaging studies (X-rays,
MRIs, CT scans), and other diagnostic tests.
5.Clinical Notes: Physicians' and nurses' notes, observations, and progress reports
from different encounters.

1.Treatment Plans: Details of prescribed treatments, procedures, surgeries, and
rehabilitation plans.
2.Vital Signs: Measurements of vital signs such as blood pressure, heart rate,
temperature, and respiratory rate.
3.Patient Visit History: Dates and summaries of previous appointments and
healthcare encounters.
4.Billing and Insurance Information: Information related to insurance coverage,
claims, and billing.
5.Doctor-Patient Communications: Records of emails, messages, and
communications between patients and healthcare providers.

Role of EHRs in consolidating patient information.
• Electronic Health Records (EHRs) play a pivotal role in consolidating and centralizing patient information, streamlining healthcare
processes, and enhancing the overall quality of patient care.
• 1. Centralized Repository: EHRs serve as a centralized and comprehensive repository for all patient-related information. They bring
together medical history, test results, diagnoses, treatment plans, prescriptions, and other clinical data from various sources into a single
digital platform.
• 2. Seamless Data Sharing: EHRs facilitate seamless data sharing among healthcare providers across different departments and locations.
This ensures that every healthcare professional involved in a patient's care has access to the most up-to-date and relevant information.
• 3. Real-Time Accessibility: EHRs provide real-time accessibility to patient information. Healthcare professionals can access EHRs instantly,
enabling them to make informed decisions quickly, especially in critical situations.
• 4. Interoperability: EHR systems are designed to be interoperable, allowing different healthcare systems and providers to exchange
information efficiently. This enables the integration of data from various sources, such as hospitals, clinics, laboratories, and pharmacies.
• 5. Continuity of Care: EHRs ensure continuity of care by providing a complete historical record of a patient's health journey. Healthcare
providers can review past diagnoses, treatments, and outcomes, enabling them to make more informed decisions about current and future
care.

• Improved Communication: EHRs enhance communication among healthcare professionals. Physicians, nurses,
specialists, and other caregivers can collaborate more effectively by accessing and updating patient information in
real-time.
• 7. Reduction of Duplication and Errors: EHRs minimize duplication of tests, procedures, and paperwork. This not
only reduces costs but also helps avoid errors due to redundant data entry.
• 8. Patient Engagement: EHRs empower patients to engage in their own healthcare. Patients can access their
records, review test results, and participate in their treatment plans, promoting shared decision-making.
• 9. Streamlined Administrative Processes: EHRs simplify administrative tasks such as appointment scheduling,
billing, and insurance processing. This allows healthcare providers to focus more on patient care.
• 10. Analytics and Reporting: EHRs enable healthcare institutions to analyze trends, outcomes, and population
health. This information supports quality improvement initiatives, resource allocation, and public health efforts.
• 11. Research and Education: EHRs provide a valuable source of data for medical research and education.
Researchers can analyze anonymized data to identify patterns, trends, and insights that contribute to medical
advancements.

Benefits: Improved patient care, data accessibility, and
clinical decision-making.
• The benefits of EHRs include:
• Improved Accessibility: Authorized healthcare professionals can access patient information in real-time, enhancing
collaboration and continuity of care.
• Accuracy and Legibility: EHRs reduce the risk of errors associated with handwritten notes and facilitate accurate
documentation.
• Efficiency: EHRs streamline administrative tasks, reduce paperwork, and enable quicker retrieval of patient information.
• Data Integration: EHRs allow for integration with various medical devices, test equipment, and software applications,
enabling seamless data sharing.
• Informed Decision-Making: EHRs provide a comprehensive view of patient health, aiding clinicians in making informed
treatment decisions.
• Patient Engagement: Patients can access their own EHRs, promoting active participation in their care and fostering better
communication with healthcare providers.
• Population Health Management: Aggregated EHR data supports public health initiatives, disease tracking, and healthcare
research.

Medical Image and Signal
Processing
• Significance of medical imaging in diagnostics and treatment.
• Medical imaging is a cornerstone of modern healthcare, playing a crucial role in both diagnostics and treatment across a wide
range of medical conditions. By providing detailed visualizations of internal structures and processes, medical imaging
technologies enable healthcare professionals to make accurate diagnoses, plan treatments, and monitor patient progress.
• 1. Early and Accurate Diagnosis: Medical imaging allows for the detection of diseases and conditions at early stages when
symptoms may not be apparent. This early diagnosis enables timely interventions and improves treatment outcomes.
• 2. Visualization of Anatomical Structures: Imaging techniques such as X-rays, CT scans, and MRI provide detailed images of
anatomical structures. These images help physicians understand the location, size, and relationships of organs and tissues.
• 3. Detection of Abnormalities: Imaging can reveal abnormalities such as tumors, fractures, and vascular conditions that
might not be visible through physical examinations alone.
• 4. Guided Treatment Planning: Imaging assists in planning surgical procedures and other treatments. Surgeons use
preoperative images to precisely navigate through complex anatomical areas, reducing risks and enhancing outcomes.

• 5. Minimally Invasive Interventions: Techniques like fluoroscopy and ultrasound aid in
minimally invasive procedures. Physicians can guide catheters, needles, and instruments to
targeted areas with real-time imaging guidance.
• 6. Monitoring Treatment Progress: Imaging allows healthcare providers to track the progress
of treatments. They can assess the response of tumors to chemotherapy, observe the healing of
fractures, and make adjustments as needed.
• 7. Interventional Radiology: Interventional radiologists use imaging to perform procedures
like angioplasty, embolization, and stent placement. These techniques are less invasive than
traditional surgery and offer quicker recovery times.
• 8. Functional Information: Imaging techniques like functional MRI (fMRI) provide information
about brain activity and function. This is valuable for understanding cognitive processes and
identifying abnormalities.

How signal processing techniques
enhance medical data analysis.
• Signal processing techniques play a critical role in enhancing the analysis of medical data, enabling healthcare professionals
and researchers to extract meaningful insights and valuable information from various types of signals. From biomedical
signals to medical images, these techniques contribute to improved diagnostics, treatment planning, and research in the field
of healthcare.
• 1. Noise Reduction: Biomedical signals such as electrocardiograms (ECGs) and electroencephalograms (EEGs) often contain
noise from various sources. Signal processing methods help filter out unwanted noise, making the signals clearer and
improving accuracy in diagnosis.
• 2. Feature Extraction: Signal processing techniques identify important features within signals that carry diagnostic or clinical
significance. These features could be peaks, troughs, or specific frequency components that reflect specific physiological
conditions.
• 3. Pattern Recognition: By analyzing patterns within signals, signal processing algorithms can identify abnormal or
characteristic patterns associated with diseases or conditions. This aids in automated disease detection and diagnosis.
• 4. Signal Enhancement: In medical imaging, signal processing methods enhance the quality of images by reducing artifacts,
enhancing contrast, and improving resolution. This leads to clearer visualization of anatomical structures and abnormalities.

• 5. Image Segmentation: Signal processing techniques enable the segmentation of medical
images to isolate regions of interest. This is essential for quantifying the size and shape of
tumors, organs, or lesions.
• 6. Quantification and Measurement: Signal processing helps measure various parameters
from medical signals and images. For example, in radiology, it quantifies bone density, blood
flow velocity, and tissue elasticity, aiding in diagnosis and treatment planning.
• 7. Data Compression: Medical data, especially imaging data, can be voluminous. Signal
processing methods allow efficient data compression without significant loss of important
information, enabling easier storage and transmission.
• 8. Real-Time Monitoring: Signal processing facilitates real-time monitoring of patient
physiological signals. This is critical in intensive care units, where timely responses to changes
in patient conditions are crucial.

Other Types of Medical Data
• Exploration of various medical data types beyond EHRs and images.
1. Genomic Data:
1. Genomic data includes information about an individual's genetic makeup.
2. It helps in understanding genetic predispositions to diseases and tailoring personalized treatment plans.
2. Patient Behavior Data:
1. Data from wearable devices and health apps capture patient behaviors like activity levels, sleep patterns, and dietary habits.
2. It contributes to preventive care, lifestyle interventions, and patient engagement.
3. Health Monitoring Devices:
1. Devices such as heart rate monitors, glucose meters, and blood pressure cuffs generate real-time physiological data.
2. Continuous monitoring assists in managing chronic conditions and detecting anomalies.
4. Biometric Data:
1. Biometric data includes fingerprints, iris scans, and voice patterns.
2. Used for secure patient identification and access control in healthcare settings.
5. Electronic Medication Dispensing Data:
1. Data from automated medication dispensers tracks patient adherence to prescribed medications.
2. It helps healthcare providers monitor medication compliance and adjust treatment plans.

Genetic data, wearable device data,
patient behavior data.
• Genetic Data:
• Genetic data involves analyzing an individual's DNA sequence and variations.
• It aids in identifying genetic predispositions to diseases, understanding hereditary conditions, and guiding precision medicine approaches.
• Wearable Device Data:
• Wearable devices collect real-time health data such as heart rate, steps taken, and sleep patterns.
• They support remote monitoring, early detection of anomalies, and promoting healthy lifestyles.
• Patient Behavior Data:
• Patient behavior data includes lifestyle choices, exercise habits, and dietary preferences.
• It assists in personalized interventions, risk assessment, and preventive care strategies.
• Biometric Data:
• Biometric data encompasses unique physiological traits like fingerprints, facial features, and voice patterns.
• Used for identity verification, access control, and patient authentication in healthcare systems.
• Environmental Data:
• Environmental data considers factors like air quality, pollution levels, and weather conditions.
• It helps correlate environmental factors with health outcomes and identify potential triggers for certain conditions.

Need for integrating diverse data
sources for comprehensive insights.
• In the rapidly evolving landscape of healthcare, integrating diverse data sources is essential to gaining a comprehensive understanding of patients' health, optimizing
treatments, and driving medical advancements.
• Holistic Patient Profiles:
• Combining electronic health records (EHRs) with genetic, wearable, and behavior data creates holistic patient profiles.
• Enables healthcare professionals to consider genetics, lifestyle, and medical history for personalized care plans.
• Early Disease Detection:
• Integration allows the detection of subtle patterns and anomalies that might not be apparent from a single data source.
• Improves early diagnosis by identifying deviations from baseline health trends.
• Predictive Analytics:
• Integrating diverse data enables predictive models for disease outcomes and treatment responses.
• Predictive analytics guide preventive measures and optimize treatment strategies.
• Personalized Medicine:
• Genetic data and patient behavior insights help tailor treatments to individual characteristics.
• Facilitates precision medicine, minimizing trial-and-error approaches.

Data Cleaning
Data cleaning, also known as data cleansing or data scrubbing, is a
critical process in healthcare analysis that involves identifying and
rectifying errors, inconsistencies, and inaccuracies present in
datasets. This ensures that the data used for analysis is accurate,
reliable, and suitable for making informed healthcare decisions.

Common data quality issues: Missing
values, outliers, inconsistent formats.
• Data quality issues can significantly impact the accuracy and reliability of healthcare analyses.
• Missing Values:
• Explanation: Missing values occur when certain data points are absent from the dataset. This can be due
to incomplete data entry, patient non-compliance, or technical issues.
• Impact: Missing values can skew statistical analyses, leading to biased results and incomplete insights.
They can hinder the accuracy of predictive models and hinder decision-making.
• Outliers:
• Explanation: Outliers are data points that significantly deviate from the rest of the dataset. They can arise
due to measurement errors, data entry mistakes, or genuinely unique cases.
• Impact: Outliers can distort statistical measures like averages and standard deviations, leading to
misleading conclusions. In healthcare, outliers might indicate rare medical conditions or data recording
errors.

• Inconsistent Formats:
• Explanation: Inconsistent data formats refer to variations in how data is represented. For example, dates may be entered in different
formats (e.g., MM/DD/YYYY vs. DD/MM/YYYY).
• Impact: Inconsistent formats complicate data analysis and aggregation. They can lead to errors in calculations and hinder effective data
comparison and integration.
• Handling Missing Values:
• Solution: Data analysts can impute missing values using methods such as mean, median, or regression imputation. Alternatively, if the
proportion of missing values is high, they might consider excluding the corresponding data points or seeking additional data sources.
• Addressing Outliers:
• Solution: Analysts can employ techniques like z-score analysis or interquartile range (IQR) to identify and handle outliers. Depending on
the context, outliers can be retained, transformed, or removed, considering their potential impact on the analysis.
• Ensuring Consistent Formats:
• Solution: Standardizing data formats ensures uniformity, simplifies analysis, and reduces errors. Using data validation rules during data
entry can prevent inconsistent formats from entering the dataset.

Importance of clean data for accurate
analysis and decision-making.
• Clean data is the foundation of accurate analysis and informed decision-making in healthcare. Ensuring that data is free from errors,
inconsistencies, and missing values is crucial for obtaining reliable insights that drive effective healthcare strategies
• Reliable Insights:
• Clean data minimizes errors and inaccuracies, leading to reliable and trustworthy insights.
• Decision-makers can confidently rely on analysis results to guide medical diagnoses, treatments, and interventions.
• Accurate Predictive Models:
• Clean data is essential for building accurate predictive models.
• Models trained on clean data produce more valid predictions, improving the accuracy of forecasting disease outcomes and treatment
responses.
• Effective Treatment Planning:
• Clinicians' base treatment plans on data-driven insights.
• Clean data ensures that treatment decisions are grounded in accurate patient information, leading to optimal care strategies.

• Minimized Bias:
• Data cleaning reduces bias introduced by errors or inconsistent data.
• Eliminating bias ensures that analyses provide a fair representation of patient populations and conditions.
• Efficient Resource Allocation:
• Clean data supports efficient allocation of healthcare resources.
• Accurate patient profiles aid in identifying areas of need, optimizing resource distribution, and enhancing patient care.
• Enhanced Patient Outcomes:
• Inaccurate data can lead to incorrect diagnoses and treatments.
• Clean data contributes to better patient outcomes by ensuring decisions are based on accurate information.
• Data-Driven Policy Making:
• Policy decisions rely on accurate data to address healthcare challenges.
• Clean data informs policy makers about trends, disparities, and healthcare needs, leading to effective public health strategies.
• Reduced Costs:
• Inaccurate data can lead to unnecessary tests, treatments, and interventions.
• Clean data helps avoid unnecessary expenditures, reducing healthcare costs.

Data Reduction Techniques
• Introduction to data reduction techniques: PCA, t-SNE.
• Data reduction techniques play a crucial role in simplifying and visualizing complex datasets while retaining
essential information. Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-
SNE) are two prominent techniques used in various fields, including healthcare, to reduce the dimensionality of
data for better analysis and visualization.
• Principal Component Analysis (PCA): PCA is a linear transformation technique that aims to reduce the
dimensionality of a dataset while preserving as much of its variance as possible. It achieves this by creating new
variables, called principal components, which are linear combinations of the original features. The first principal
component captures the most variance in the data, followed by subsequent components in decreasing order. PCA
is particularly useful when dealing with high-dimensional data, such as medical imaging or genetic data.
• t-Distributed Stochastic Neighbor Embedding (t-SNE): t-SNE is a nonlinear dimensionality reduction technique
that emphasizes maintaining the pairwise similarities between data points in a lower-dimensional space. It is
particularly effective at capturing local structures and clusters within data. t-SNE focuses on retaining the
relationships between data points, making it suitable for visualization and exploratory analysis, especially when
dealing with complex datasets.

Purpose of reducing dimensionality
while preserving essential information.
• Reducing dimensionality while preserving essential information
serves multiple purposes across various fields, including data
analysis, machine learning, and visualization. The goal is to simplify
complex datasets without sacrificing critical insights or patterns.

How PCA transforms correlated
variables into orthogonal components.
• Principal Component Analysis (PCA) is a dimensionality reduction technique that transforms the original features of a dataset into a new set of
uncorrelated variables called principal components. This transformation is achieved through linear combinations of the original variables,
effectively reducing the dimensionality of the data while retaining as much variance as possible. PCA has the unique property of being able to
transform correlated variables into orthogonal (uncorrelated) components.
• Covariance Matrix Calculation:
• PCA begins by calculating the covariance matrix of the original data. The covariance matrix represents the relationships and dependencies
between the different variables in the dataset.
• Eigenvector-Eigenvalue Decomposition:
• The next step involves performing an eigenvector-eigenvalue decomposition on the covariance matrix. Eigenvectors are unit vectors that
represent the directions along which the data varies the most, and eigenvalues indicate the amount of variance explained by each
eigenvector.
• Selection of Principal Components:
• The eigenvectors with the highest eigenvalues (most variance) are selected as the principal components. These eigenvectors represent the
new orthogonal axes in the transformed space.

• Orthogonal Transformation:
• The original variables are projected onto the new orthogonal axes defined by the selected principal components. The
resulting projections are the transformed variables, or principal component scores.
• Uncorrelated Principal Components:
• The crucial property of PCA is that the principal components are orthogonal to each other. This means that they are
uncorrelated, and the correlation between any pair of principal components is zero.
• Variance Preservation:
• While the principal components are uncorrelated, they are ordered by the amount of variance they explain in the
original data. The first principal component captures the most variance, followed by subsequent components.
• Dimensionality Reduction:
• The dimensions of the data can be reduced by keeping only a subset of the principal components that explain a
significant portion of the total variance. This reduction simplifies the data representation while preserving essential
information.

Feature Selection
• Definition of feature selection in the context of healthcare.
• Feature selection refers to the process of identifying and selecting a subset of relevant and informative features
(variables) from a larger set of potential features in a healthcare dataset.
• In other words, it involves choosing the most influential attributes that have a significant impact on the outcome
of interest.
• The goal of feature selection in healthcare is to improve the efficiency, effectiveness, and interpretability of data
analysis and modeling tasks while maintaining or enhancing the quality of results.
• Feature selection is particularly important in healthcare due to the complexity and dimensionality of medical
datasets.
• It enables healthcare professionals and researchers to focus on the most relevant information, reduce noise and
redundancy, and enhance the accuracy of predictive models, diagnostic tools, and treatment recommendations.
• By selecting the right features, healthcare practitioners can gain insights, identify patterns, and make informed
decisions that contribute to better patient care and medical advancements.

Role in improving model performance
and interpretability.
• Feature selection plays a pivotal role in improving both the performance and interpretability of models in the field of healthcare. It enhances the efficiency and effectiveness of data analysis while making the results more understandable
and actionable.
• 1. Improved Model Performance:
• Reduced Overfitting: Feature selection helps prevent overfitting by eliminating irrelevant or noisy features. Overfitting occurs when a model captures noise in the data, leading to poor generalization on new data.
• Enhanced Generalization: By focusing on the most informative features, models become more generalizable to new, unseen data. They capture underlying patterns rather than memorizing the training data.
• Optimized Complexity: A smaller set of features leads to simpler models, reducing the risk of model complexity that can hinder performance on new data.
• 2. Enhanced Interpretability:
• Clearer Insights: With fewer features, the relationships between input variables and outcomes become clearer. Healthcare professionals can better understand how specific factors influence patient outcomes or disease risks.
• Intuitive Explanations: Simplified models resulting from feature selection are easier to explain to both healthcare practitioners and patients. Transparent explanations foster trust in the model's decisions.
• Clinical Relevance: Selected features are likely to be those that are clinically relevant and meaningful to healthcare experts, aiding in the interpretation of model results.
• .

• 3. Reduced Noise and Redundancy:
• Noise Removal: Eliminating irrelevant or noisy features reduces the impact of irrelevant fluctuations in the data that might mislead the model.
• Redundancy Reduction: When multiple features convey similar information, feature selection helps retain the most informative ones, reducing redundancy.
• 4. Faster Model Training and Inference:
• Efficient Processing: Fewer features lead to faster model training and predictions, making the entire workflow more time-efficient.
• Real-Time Applications: In healthcare settings where quick decisions are critical, feature selection enables models to provide timely insights and recommendations.
• 5. Ethical and Regulatory Compliance:
• Sensitive Information: Feature selection can help exclude sensitive patient data, ensuring compliance with privacy regulations like HIPAA.
• Transparency: Removing irrelevant features avoids unintentional bias and contributes to the transparency of model decisions

Methods for selecting relevant features:
Filter, wrapper, embedded approaches.
• In the context of healthcare data analysis and modeling, selecting relevant features is crucial for building accurate and interpretable models.
Three common approaches for feature selection are filter, wrapper, and embedded methods. Each approach has its advantages and limitations.
• 1. Filter Approach:
• Concept: The filter approach involves evaluating the relevance of features based on their individual characteristics, such as correlation with
the target variable or statistical tests. Features are ranked or scored independently of the chosen model.
• Advantages:
• Computationally efficient, as it doesn't involve model training.
• Useful for preliminary feature selection to identify potentially informative attributes.
• Limitations:
• Ignores interactions among features.
• May select features that are individually relevant but not collectively meaningful.

• 2. Wrapper Approach:
• Concept: The wrapper approach selects features by treating the feature selection process as part of the
model-building process. It uses a specific machine learning algorithm to train and evaluate the model
iteratively with different subsets of features.
• Advantages:
• Considers interactions among features and their impact on model performance.
• Tailored to the specific algorithm used, leading to potentially better-selected features for that model.
• Limitations:
• Computationally intensive and time-consuming, especially with large datasets.
• Prone to overfitting if not properly validated.

• 3. Embedded Approach:
• Concept: The embedded approach incorporates feature selection into the model training process itself.
During model training, certain algorithms automatically perform feature selection based on their
importance for model performance.
• Advantages:
• Efficient integration of feature selection and model building.
• Takes advantage of the inherent feature importance estimation of certain algorithms.
• Limitations:
• Limited to models that support embedded feature selection, potentially reducing flexibility in model
choice.
• May not provide a global view of feature importance beyond the specific model used.

Data Management ,Data Collection,Data Pre-processing for AI&ML in Healthcare

More Related Content

Similar to Data Management ,Data Collection,Data Pre-processing for AI&ML in Healthcare (20)

Recently uploaded (20)

Data Management ,Data Collection,Data Pre-processing for AI&ML in Healthcare