Supervised Machine Learning Methods Research Articles

Two-stage procedure for transportation mode detection based on sighting data

The data required for transportation applications can be retrieved from mobile phones without the necessity of additional infrastructure. Thus, we propose a procedure that involves two stages – data preprocessing and transportation mode detection – for detecting the transportation mode (i.e., car and bus) on the basis of sighting data. In the data preprocessing stage, two detection rules are used for eliminating oscillations that occur when a mobile phone intermittently switches between cell towers instead of connecting to the nearest cell tower. In the transportation mode detection stage, two supervised machine learning methods, namely support vector machine (SVM) and a deep neural network (DNN), are used to detect transportation modes. Experimental results indicated SVM achieved a higher accuracy (96.49%) in transport mode detection than did the DNN (69.65%) during peak hours. Moreover, travel time and starting time of a trip were identified as critical features affecting the accuracy of transportation mode detection.

A big data analytics framework for determining the travel destination preferences of Indian tourists

The growth of technology and social media websites has increased the potential to online explore different products and places around the globe. While online websites are primarily responsible for the generation of large amounts of data, this big data may be beneficial to other users provided the proper decision pattern can be analyzed. This work is focusing on the big data from social media to determine the travel destination preferences for Indian tourists. The analysis of online tourism reviews is beneficial to both tourists and businesses in tourist countries. Tourists can analyze all the required aspects prior to traveling and businesses in the destination country can enhance their products. The study aims to analyze the online tourist reviews using supervised machine learning methods (decision tree, k-nearest neighbor, back propagation neural networks and support vector machine) and ensemble learning in order to ascertain the travel preferences of Indian tourists visiting other countries. For the research experiments, significant travel data histories of tourists for the five destination places (Dubai, Indonesia, Malaysia, Thailand and Singapore) are extracted from TripAdvisor. TripAdvisor is a worldwide popular tourism website that provides access to consumers to share their travel experiences. From the selected five destination places, the preferences of Indian tourists are analyzed for the factors of travel & destination comfort, hotel facilities, food quality and attractions of the place. The analysis results of the proposed recommendation system indicate the determination of precise suggestions for Indian tourists traveling to other countries.

District heating planning with focus on solar energy and heat pump using GIS and the supervised learning method: Case study of Gaziantep, Turkey

In the present context, the global concern on energy consumption and management have been significantly increased due to the environmental issues, such as global warming and greenhouse gas emission. The heating supply is one of the most energy-intensive applications at present, and this study presents an effective model for planning and utilizing a district heating system. Further, the model is applied to a province in Turkey to fulfill environmental, technical, and economic goals. In the first step, indices have been used, including demographics, efficiency of the buildings, and the number of households, to predict the required heating load by support vector regression (SVR) as a supervised machine learning method until 2030. The heat energy demand would be increased by 9% in 2030 compared to 2020. Thereafter, most suitable regions are evaluated to establish district heating systems based on geographic information system (GIS). The classification of Gaziantep province shows that more than 70% of the area is suitable for establishing a solar-based district heating system. The center of the province including Shahinbey, Sehitkamil, and Araban, is the highest priority to integrate a solar energy system into the existing energy system to maximize its share of the energy system. Therefore, in this research, five general scenarios including different combinations of heat pump (HP), solar thermal (ST), photovoltaic (PV) system, battery (BT), and heat storage (HS) are defined and analyzed to determine the most effective scenario, in terms of economic and environmental aspects. Finally, results show all scenarios could reduce the CO2 emissions; however, the combination of ST and HP has the least costs due to the 21.8% reduction in the total primary energy (TPE) supply compared to BAU. Applying solar energy with a heat pump (S5) leads a 37% reduction in CO2 emissions compared to BAU. Overall, the minimum emissions is belonged to scenario 5, including solar heat pump and storage. Moreover, the effects of parameters such as carbon taxes, technological advancements, and electricity prices are evaluated by to sensitivity analysis to confirm the reliability of the results.

Applying Supervised Machine Learning Methods to Determine the Atomicity of Requirements for Complex Technical Systems

The article is devoted to the problem of determining the atomicity of requirements for complex technical systems. The purpose of this article is to apply supervised machine learning methods, namely classification, to determine the atomicity of requirements. It is assumed that feature engineering based on the linguistic features of requirements stateВ­ments will make it possible to classify requirements into atomic and non-atomic ones with high accuracy. The article describes the use of the following methods for generating features: fastText, BERT, customFeature. Requirements are classified using a logistic classifier, decision trees, random forest, and gradient boosting. The best combination of methods turned out to be using customFeature with gradient boosting.

A data-centric weak supervised learning for highway traffic incident detection

Using the data from loop detector sensors for near-real-time detection of traffic incidents on highways is crucial to averting major traffic congestion. While recent supervised machine learning methods offer solutions to incident detection by leveraging human-labeled incident data, the false alarm rate is often too high to be used in practice. Specifically, the inconsistency in the human labeling of the incidents significantly affects the performance of supervised learning models. To that end, we focus on a data-centric approach to improve the accuracy and reduce the false alarm rate of traffic incident detection on highways. We develop a weak supervised learning workflow to generate high-quality training labels for the incident data without the ground truth labels, and we use those generated labels in the supervised learning setup for final detection. This approach comprises three stages. First, we introduce a data preprocessing and curation pipeline that processes traffic sensor data to generate high-quality training data through leveraging labeling functions, which can be domain knowledge-related or simple heuristic rules. Second, we evaluate the training data generated by weak supervision using three supervised learning models—random forest, k-nearest neighbors, and a support vector machine ensemble—and long short-term memory classifiers. The results show that the accuracy of all of the models improves significantly after using the training data generated by weak supervision. Third, we develop an online real-time incident detection approach that leverages the model ensemble and the uncertainty quantification while detecting incidents. Overall, we show that our proposed weak supervised learning workflow achieves a high incident detection rate (0.90) and low false alarm rate (0.08).

Engagement estimation of the elderly from wild multiparty human–robot interaction

AbstractThe use of social robots in healthcare systems or nursing homes to assist the elderly and their caregivers will be becoming common, where robots' understanding of engagement of the elderly is important. Traditional engagement estimation (EE) often requires expert involvement in a controlled dyadic interaction environment. In this article, we propose a supervised machine learning method to estimate the engagement state of the elderly in a multiparty human–robot interaction (HRI) scenario from the real‐world video recording as input. The method is built upon the basic concept of engagement in geriatric psychiatry and HRI video representations. It adapts pretrained models to extract behavior, affective, and visual signals to form the multi‐modal features. These features are then fed into a neural network made of a self‐attention mechanism and average pooling for individual learning, a graph attention network for group learning and a fully connected layer to estimate the engagement. We tested the proposed method using 43 wild multiparty elderly robot interaction (ERI) videos. The experimental results show that our method is capable of detecting the key participants and estimating the engagement state of the elderly effectively. Also our study demonstrates the signals from side‐participants in the main interaction group considerably contribute to the EE of the elderly in the multiparty ERI.

A Novel Multi-Sensor Data-Driven Approach to Source Term Estimation of Hazardous Gas Leakages in the Chemical Industry

Source term estimation (STE) is crucial for understanding and addressing hazardous gas leakages in the chemical industry. Most existing methods basically use an atmospheric transport and dispersion (ATD) model to predict the concentrations of hazardous gas leakages from different possible sources, compare the predicted results with multi-sensor data, and use the deviations to search and derive information on the real sources of leakages. Although performing well in principle, complicated computations and the associated computer time often make these methods difficult to apply in real time. Recently, many machine learning methods have also been proposed for the purpose of STE. The idea is to build offline a machine-learning-based STE model using data generated with a high-fidelity ATD model and then apply the machine learning model to multi-sensor data to perform STE in real time. The key to the success of a machine-learning-based STE is that the machine-learning-based STE model has to cover all possible scenarios of concern, which is often difficult in practice because of unpredictable environmental conditions and the inherent robust problems with many supervised machine learning methods. In order to address challenges with the existing STE methods, in the present study, a novel multi-sensor data-driven approach to STE of hazardous gas leakages is proposed. The basic idea is to establish a multi-sensor data-driven STE model from historical multi-sensor observations that cover the situations known as the independent hazardous-gas-leakage scenarios (IHGLSs) in a chemical industry park of concern. Then the established STE model is applied to online process multi-sensor data and perform STE for the chemical industry park in real time. The new approach is based on a rigorous analysis of the relationship between multi-sensor data and sources of hazardous gas leakages and derived using advanced data science, including unsupervised multi-sensor data clustering and analysis. As an example of demonstration, the proposed approach is applied to perform STE for hazardous gas-leakage scenarios wherein a Gaussian plume model can be used to describe the atmospheric transport and dispersion. Because of no need of ATD-model-based online optimization and supervised machine learning, the new approach can potentially overcome many problems with existing methods and enable STE to be literally applied in engineering practice.

Rational Designing of Bimetallic/Trimetallic Hydrogen Evolution Reaction Catalysts Using Supervised Machine Learning.

Cost-efficient electrocatalysts to replace precious platinum group metals- (PGMs-) based catalysts for the hydrogen evolution reaction (HER) carry significant potential for sustainable energy solutions. Machine learning (ML) methods have provided new avenues for intelligent screening and predicting efficient heterogeneous catalysts in recent years. We coalesce density functional theory (DFT) and supervised ML methods to discover earth-abundant active heterogeneous NiCoCu-based HER catalysts. An intuitive generalized microstructure model was designed to study the adsorbate's surface coverage and generate input features for the ML process. The study utilizes optimized eXtreme Gradient Boost Regression (XGBR) models to screen NiCoCu alloy-based catalysts for HER. We show that the most active HER catalysts can be screened from an extensive set of catalysts with this approach. Therefore, our approach can provide an efficient way to discover novel heterogeneous catalysts for various electrochemical reactions.

Machine learning for risk stratification in kidney disease.

Risk stratification for chronic kidney is becoming increasingly important as a clinical tool for both treatment and prevention measures. The goal of this review is to identify how machine learning tools contribute and facilitate risk stratification in the clinical setting. The two key machine learning paradigms to predictively stratify kidney disease risk are genomics-based and electronic health record based approaches. These methods can provide both quantitative information such as relative risk and qualitative information such as characterizing risk by subphenotype. The four key methods to stratify chronic kidney disease risk are genomics, multiomics, supervised and unsupervised machine learning methods. Polygenic risk scores utilize whole genome sequencing data to generate an individual's relative risk compared with the population. Multiomic methods integrate information from multiple biomarkers to generate trajectories and prognostic different outcomes. Supervised machine learning methods can directly utilize the growing compendia of electronic health records such as laboratory results and notes to generate direct risk predictions, while unsupervised machine learning methods can cluster individuals with chronic kidney disease into subphenotypes with differing approaches to care.

Analysis of Different Supervised Machine Learning Methods for Accelerometer-Based Alcohol Consumption Detection from Physical Activity

Analysis of Different Supervised Machine Learning Methods for Accelerometer-Based Alcohol Consumption Detection from Physical Activity

Cognitive load detection using circulant singular spectrum analysis and Binary Harris Hawks Optimization based feature selection

Cognitive load detection during the mental assignment of neural activity is necessary because it helps to understand the brain’s response to stimuli. An electroencephalogram (EEG) can be used to identify cognitive load during mental arithmetic activities. EEG data was collected from public databases such as the mental arithmetic task (MAT) and simultaneous task workload (STEW). In this manuscript, short-term EEG signals were used to detect cognitive load. Circulant singular spectrum analysis (C-SSA) was used to decompose the EEG signals into intrinsic mode functions (IMF’s). After that, we extract the entropy based features from the IMF’s. For feature selection, optimization algorithms were used, namely Binary Grey Wolf Optimization (BGWO), Binary Harris Hawks Optimization (BHH0), and Binary Differential Evolution (BDE). Furthermore, supervised machine learning methods, namely K-nearest neighbour (KNN) and support vector machine (SVM), were employed to classify selected features based on performance metrics like accuracy (ACC %), sensitivity (SEN), specificity (SPE), precision (PRE), and F-Score (F-S). C-SSA for decomposition and optimization algorithms for feature selection are novel techniques for cognitive load detection. The proposed BHHO-KNN technique achieves the best classification accuracy of 96.88% and 95.28%, individually, for the STEW and MAT datasets. The conclusions of the experiments showed that the proposed technique is more precise at detecting cognitive load compared to existing methods.

Deep Temporal Iterative Clustering for Satellite Image Time Series Land Cover Analysis

The extensive amount of Satellite Image Time Series (SITS) data brings new opportunities and challenges for land cover analysis. Many supervised machine learning methods have been applied in SITS, but the labeled SITS samples are time- and effort-consuming to acquire. It is necessary to analyze SITS data with an unsupervised learning method. In this paper, we propose a new unsupervised learning method named Deep Temporal Iterative Clustering (DTIC) to deal with SITS data. The proposed method jointly learns a neural network’s parameters and the resulting features’ cluster assignments, which uses a standard clustering algorithm, K-means, to iteratively cluster the features produced by the feature extraction network and then uses the subsequent assignments as supervision to update the network’s weights. We apply DTIC to the unsupervised training of neural networks on both SITS datasets. Experimental results demonstrate that DTIC outperforms the state-of-the-art K-means clustering algorithm, which proves that the proposed approach successfully provides a novel idea for unsupervised training of SITS data.

Morphological Dynamics-Based Anomaly Detection Towards In Situ Layer-Wise Certification for Directed Energy Deposition Processes

Abstract The process uncertainty induced quality issue remains the major challenge that hinders the wider adoption of additive manufacturing (AM) technology. The defects occurred significantly compromise structural integrity and mechanical properties of fabricated parts. Therefore, there is an urgent need in fast, yet reliable AM component certification. Most finite element analysis related methods characterize defects based on the thermomechanical relationships, which are computationally inefficient and cannot capture process uncertainty. In addition, there is a growing trend in data-driven approaches on characterizing the empirical relationships between thermal history and anomaly occurrences, which focus on modeling an individual image basis to identify local defects. Despite their effectiveness in local anomaly detection, these methods are quite cumbersome when applied to layer-wise anomaly detection. This paper proposes a novel in situ layer-wise anomaly detection method by analyzing the layer-by-layer morphological dynamics of melt pools and heat affected zones (HAZs). Specifically, the thermal images are first preprocessed based on the g-code to assure unified orientation. Subsequently, the melt pool and HAZ are segmented, and the global and morphological transition metrics are developed to characterize the morphological dynamics. New layer-wise features are extracted, and supervised machine learning methods are applied for layer-wise anomaly detection. The proposed method is validated using the directed energy deposition (DED) process, which demonstrates superior performance comparing with the benchmark methods. The average computational time is significantly shorter than the average build time, enabling in situ layer-wise certification and real-time process control.

Machine Learning Model-Based Ice Cover Forecasting for a Vital Waterway in Large Lakes

The St. Marys River is a key waterway that supports the navigation activities in the Laurentian Great Lakes. However, high year-to-year fluctuations in ice conditions pose a challenge to decision making with respect to safe and effective navigation, lock operations, and ice breaking operations. The capability to forecast the ice conditions for the river system can greatly aid such decision making. Small-scale features and complex physics in the river system are difficult to capture by process-based numerical models that are often used for lake-wide applications. In this study, two supervised machine learning methods, the Long Short-Term Memory (LSTM) model and the Extreme Gradient Boost (XGBoost) algorithm are applied to predict the ice coverage on the St. Marys River for short-term (7-day) and sub-seasonal (30-day) time scales. Both models are trained using 25 years of meteorological data and select climate indices. Both models outperform the baseline forecast in the short-term applications, but the models underperform the baseline forecast in the sub-seasonal applications. The model accuracies are high in the stable season, while they are lower in the freezing and melting periods when ice conditions can change rapidly. The errors of the predicted ice-on/ice-off date lie within 2–5 days.

Classification of COVID-19 from tuberculosis and pneumonia using deep learning techniques

Deep learning provides the healthcare industry with the ability to analyse data at exceptional speeds without compromising on accuracy. These techniques are applicable to healthcare domain for accurate and timely prediction. Convolutional neural network is a class of deep learning methods which has become dominant in various computer vision tasks and is attracting interest across a variety of domains, including radiology. Lung diseases such as tuberculosis (TB), bacterial and viral pneumonias, and COVID-19 are not predicted accurately due to availability of very few samples for either of the lung diseases. The disease could be easily diagnosed using X-ray or CT scan images. But the number of images available for each of the disease is not as equally as other resulting in imbalance nature of input data. Conventional supervised machine learning methods do not achieve higher accuracy when trained using a lesser amount of COVID-19 data samples. Image data augmentation is a technique that can be used to artificially expand the size of a training dataset by creating modified versions of images in the dataset. Data augmentation helped reduce overfitting when training a deep neural network. The SMOTE (Synthetic Minority Oversampling Technique) algorithm is used for the purpose of balancing the classes. The novelty in this research work is to apply combined data augmentation and class balance techniques before classification of tuberculosis, pneumonia, and COVID-19. The classification accuracy obtained with the proposed multi-level classification after training the model is recorded as 97.4% for TB and pneumonia and 88% for bacterial, viral, and COVID-19 classifications. The proposed multi-level classification method produced is ~8 to ~10% improvement in classification accuracy when compared with the existing methods in this area of research. The results reveal the fact that the proposed system is scalable to growing medical data and classifies lung diseases and its sub-types in less time with higher accuracy.Graphical

Addressing Binary Classification over Class Imbalanced Clinical Datasets Using Computationally Intelligent Techniques.

Nowadays, healthcare is the prime need of every human being in the world, and clinical datasets play an important role in developing an intelligent healthcare system for monitoring the health of people. Mostly, the real-world datasets are inherently class imbalanced, clinical datasets also suffer from this imbalance problem, and the imbalanced class distributions pose several issues in the training of classifiers. Consequently, classifiers suffer from low accuracy, precision, recall, and a high degree of misclassification, etc. We performed a brief literature review on the class imbalanced learning scenario. This study carries the empirical performance evaluation of six classifiers, namely Decision Tree, k-Nearest Neighbor, Logistic regression, Artificial Neural Network, Support Vector Machine, and Gaussian Naïve Bayes, over five imbalanced clinical datasets, Breast Cancer Disease, Coronary Heart Disease, Indian Liver Patient, Pima Indians Diabetes Database, and Coronary Kidney Disease, with respect to seven different class balancing techniques, namely Undersampling, Random oversampling, SMOTE, ADASYN, SVM-SMOTE, SMOTEEN, and SMOTETOMEK. In addition to this, the appropriate explanations for the superiority of the classifiers as well as data-balancing techniques are also explored. Furthermore, we discuss the possible recommendations on how to tackle the class imbalanced datasets while training the different supervised machine learning methods. Result analysis demonstrates that SMOTEEN balancing method often performed better over all the other six data-balancing techniques with all six classifiers and for all five clinical datasets. Except for SMOTEEN, all other six balancing techniques almost had equal performance but moderately lesser performance than SMOTEEN.

Using supervised machine learning for large‐scale classification in management research: The case for identifying artificial intelligence patents

AbstractResearch SummaryResearchers increasingly use unstructured text data to construct quantitative variables for analysis. This goal has traditionally been achieved using keyword‐based approaches, which require researchers to specify a dictionary of keywords mapped to the theoretical concepts of interest. However, recent machine learning (ML) tools for text classification and natural language processing can be used to construct quantitative variables and to classify unstructured text documents. In this paper, we demonstrate how to employ ML tools for this purpose and discuss one application for identifying artificial intelligence (AI) technologies in patents. We compare and contrast various ML methods with the keyword‐based approach, demonstrating the advantages of the ML approach. We also leverage the classification outcomes generated by ML models to demonstrate general patterns of AI technological innovation development.Managerial SummaryText‐based documents offer a wealth of information for researchers and business analysts. However, researchers often need to find a way to classify these documents to use in subsequent research projects. In this paper, we demonstrate how supervised ML methods can be used to automate the process of classifying textual documents into pre‐defined categories or groups. We provide an overview of when such techniques may be used in comparison to other methods, and the considerations and tradeoffs associated with each method. We apply these methods to identify AI‐based technologies from all patents in the United States, based on patent abstract text. This allows us to show interesting patterns of AI innovation development in the United States. We also provide the code and data used in this paper for future research.

The efficacy of canagliflozin in diabetes subgroups stratified by data-driven clustering or a supervised machine learning method: a post hoc analysis of canagliflozin clinical trial data.

Data-driven diabetes subgroups have shown distinct clinical characteristics and disease progression, although there is a lack of evidence that this information can guide clinical decisions. We aimed to investigate whether diabetes subgroups, identified by data-driven clustering or supervised machine learning methods, respond differently to canagliflozin. We pooled data from five randomised, double-blinded clinical trials of canagliflozin at an individual level. We applied the coordinates from the All New Diabetics in Scania (ANDIS) study to form four subgroups: mild age-related diabetes (MARD); severe insulin-deficient diabetes (SIDD); mild obesity-related diabetes (MOD) and severe insulin-resistant diabetes (SIRD). Machine learning models for HbA1c lowering (ML-A1C) and albuminuria progression (ML-ACR) were developed. The primary efficacy endpoint was reduction in HbA1c at 52 weeks. Concordance of a model was defined as the difference between predicted HbA1c and actual HbA1c decline less than 3.28 mmol/mol (0.3%). The decline in HbA1c resulting from treatment was different among the four diabetes clusters (pinteraction=0.004). In MOD, canagliflozin showed a robust glucose-lowering effect at week 52, compared with other drugs, with least-squares mean of HbA1c decline [95% CI] being 6.6 mmol/mol (4.1, 9.2) (0.61% [0.38, 0.84]) for sitagliptin, 7.1 mmol/mol (4.7, 9.5) (0.65% [0.43, 0.87]) for glimepiride, and 9.8 mmol/mol (9.0, 10.5) (0.90% [0.83, 0.96]) for canagliflozin. This superiority persisted until 104 weeks. The proportion of individuals who achieved HbA1c <53 mmol/mol (<7.0%) was highest in sitagliptin-treated individuals with MARD but was similar among drugs in individuals with MOD. The ML-A1C model and the cluster algorithm showed a similar concordance rate in predicting HbA1c lowering (31.5% vs 31.4%, p=0.996). Individuals were divided into high-risk and low-risk groups using ML-ACR model according to their predicted progression risk for albuminuria. The effect of canagliflozin vs placebo on albuminuria progression differed significantly between the high-risk (HR 0.67 [95% CI 0.57, 0.80]) and low-risk groups (HR 0.91 [0.75, 1.11]) (pinteraction=0.016). Data-driven clusters of individuals with diabetes showed different responses to canagliflozin in glucose lowering but not renal outcome prevention. Canagliflozin reduced the risk of albumin progression in high-risk individuals identified by supervised machine learning. Further studies with larger sample sizes for external replication and subtype-specific clinical trials are necessary to determine the clinical utility of these stratification strategies in sodium-glucose cotransporter 2 inhibitor treatment. The application for the clinical trial data source is available on the YODA website ( http://yoda.yale.edu/ ).

Novel deep neural network framework for biomedical named entity recognition

With the dramatic improvements in the field of bio-informatics, extracting information from text and analyzing the association between the entities has received more attention in the past few years. Entity Recognition (ER) meant to extract and recognize the entities from any text. Biomedical Named Entity Recognition (BNER) gets more and more attention from the researchers since it is a fundamental task in biomedical information extraction. Various methods has been proposed to perform the task of BioNER. Different kind of approaches are dictionary based, rule based approaches, traditional machine learning approaches that combines supervised and unsupervised methods and neural network based approach. The state-of-art systems previously adopted various supervised machine learning methods Hidden Markov Models (HMMs), Maximum Entropy Markov Models (MEMMs), Support vector machines(SVM),Structural Support Vector Machines(SSVMS),Conditional Random Fields(CRF) to derive semantic and syntatic features from annotated datasets. However, CRF is one of the most successful method used for NER and it has obtained finest result because of the robustness and ability for sequence lebelling task. Recently, studies have demonstrated the application of deep learning based approaches for biomedical named entity recognition (BioNER) and shown promising results.

An Assessment and Analysis Model of Psychological Health of College Students Based on Convolutional Neural Networks.

Psychological health assessment and psychological problem identification essentially belong to problems of pattern recognition or nonlinear classification; its system contains complex nonlinear interactions among various factors, having basic characteristics of multivariable, multilevel, and strong coupling. An important problem in the field of artificial intelligence solved by convolutional neural networks (CNN) is to simplify complex problems, minimize the number of parameters, and thus greatly improve the algorithm's performance. Therefore, CNN has outstanding advantages in establishing the assessment and analysis model of college students' psychological health. This study determined the psychological health standards of college students, selected measurement tools for college students' psychological state, elaborated the principles of psychological assessment based on text information, performed the sample set data establishment and data processing of the assessment and analysis model of psychological health, conducted network establishment, training, and simulation, carried out a case experiment and its result analysis, explored the cause analysis of college students' psychological health problems, and finally discussed the prevention and intervention of college students' psychological problems. The study results show that the input and output of the CNN-based assessment and analysis model of college students' psychological health are their evaluation data and assessment results, respectively, and the optimal hyperparameters of the model are determined through fold cross-validation analysis to improve the model's over-fitting problem. After the training is completed, the model can predict the changes in college students' psychological state in the future through the psychological test data. The CNN uses supervised machine learning method to construct an assessment and analysis model of college students' psychological health, and establishes the mapping relationship between college students' personal background and their psychological health. The network error continuously adjusts network connection weight according to gradient descent algorithm to minimize its error, so that the convolutional layer and the pooling layer can learn the optimized feature expression of the input data.