Classification-based Approaches Research Articles

Comparative evaluation of performances of algae indices, pixel- and object-based machine learning algorithms in mapping floating algal blooms using Sentinel-2 imagery

One of the main threats to freshwater resources is pollution from anthropogenic activities such as rapid urbanization and excessive agricultural nutrient runoff. Remote sensing technologies have been effectively used in monitoring and mapping rapid changes in the marine environment and assessing the overall health of freshwater ecosystems. The main goal of this study is to comparatively evaluate the performance of index-based and classification-based approaches in mapping dense floating algal blooms observed in Lake Burdur using Sentinel-2 imagery. For index-based mapping, algae-specific indices, namely the Floating Algae Index (FAI), Adjusted Floating Algae Index, Surface Algal Blooms Index (SABI), and Algal Blooms Detection Index (ABDI), were used. At the same time, pixel- and object-based Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Long Short-Term Memory Network (LSTM) were utilized for classification-based algal mapping. For this purpose, seven Sentinel-2 images, selected through time series analysis performed on the Google Earth Engine platform, were used as the primary dataset in the application. The results show that high-density floating algae formations can be detected over 99% by both indices and classification-based approaches, whereas pixel-based classification is more successful in mapping low-density algal blooms. When two-class thematic maps representing water and floating algae classes were considered, the maps produced by index-based FAI using an appropriate threshold value and the classification-based RF algorithm reached an overall accuracy of over 99%. The highest algae density in the lake was observed on July 13, 2021, and was determined to be effective in ~ 45 km2 of the lake’s surface.

Learning phase transitions from regression uncertainty: a new regression-based machine learning approach for automated detection of phases of matter

Abstract For performing regression tasks involved in various physics problems, enhancing the precision or equivalently reducing the uncertainty of regression results is undoubtedly one of the central goals. Here, somewhat surprisingly, the unfavorable regression uncertainty in performing the regression tasks of inverse statistical problems is found to contain hidden information concerning the phase transitions of the system under consideration. By utilizing this hidden information, a new unsupervised machine learning approach was developed in this work for automated detection of phases of matter, dubbed learning from regression uncertainty. This is achieved by revealing an intrinsic connection between regression uncertainty and response properties of the system, thus making the outputs of this machine learning approach directly interpretable via conventional notions of physics. It is demonstrated by identifying the critical points of the ferromagnetic Ising model and the three-state clock model, and revealing the existence of the intermediate phase in the six-state and seven-state clock models. Comparing to the widely-used classification-based approaches developed so far, although successful, their recognized classes of patterns are essentially abstract, which hinders their straightforward relation to conventional notions of physics. These challenges persist even when one employs the state-of-the-art deep neural networks (NNs) that excel at classification tasks. In contrast, with the core working horse being an NN performing regression tasks, our new approach is not only practically more efficient, but also paves the way towards intriguing possibilities for unveiling new physics via machine learning in a physically interpretable manner.

Real-time wear rate prediction and analysis: Gradient-weighted class activation mapping (Grad-CAM) in 1D convolutional neural network bridges experiments and neural networks

In this study, we propose a one-dimensional convolutional neural network (1D-CNN) for predicting wear rates using friction coefficient data. The model achieves high prediction accuracy for both training and testing sets, surpassing other machine learning methods. To ensure model reliability, we employ the Grad-CAM method, calculating importance scores correlating well with wear severity assessed by surface roughness (Ra) and wear track topography. The 1D-CNN model promises a precise, quantitative assessment of wear severity compared to traditional classification-based approaches. Furthermore, the model exhibits robust generalization abilities and potential as a base model for predicting wear rates of other materials, broadening its applicability in the tribology field.

Rotated Object Detection with Circular Gaussian Distribution

Rotated object detection is a challenging task due to the difficulties of locating the rotated objects and separating them effectively from the background. For rotated object prediction, researchers have explored numerous regression-based and classification-based approaches to predict a rotation angle. However, both paradigms are constrained by some flaws that make it difficult to accurately predict angles, such as multi-solution and boundary issues, which limits the performance upper bound of detectors. To address these issues, we propose a circular Gaussian distribution (CGD)-based method for angular prediction. We convert the labeled angle into a discrete circular Gaussian distribution spanning a single minimal positive period, and let the model predict the distribution parameters instead of directly regressing or classifying the angle. To improve the overall efficiency of the detection model, we also design a rotated object detector based on CenterNet. Experimental results on various public datasets demonstrated the effectiveness and superior performances of our method. In particular, our approach achieves better results than state-of-the-art competitors, with improvements of 1.92% and 1.04% in terms of AP points on the HRSC2016 and DOTA datasets, respectively.

Optimized real-time parking management framework using deep learning

Today’s modern world has seen massive population growth in the past few decades, causing the automobile industry to expand rapidly. The exponential increase has disrupted the seamless flow of traffic globally and has fashioned significant problems such as traffic congestion and its mismanagement. To tackle these issues, some hardware and data-driven-based solutions have been previously proposed. The hardware-based solutions utilize different types of sensors installed at relevant sites to monitor the status of parking slots which makes them less scalable, complex, costly to install and maintain. In contrast, the data-driven solutions utilize the already available infrastructure of surveillance cameras installed at parking lots, thus overcoming the limitations of sensor-based solutions. However, currently only classification-based approaches are prevalent and adopted to monitor the status of the parking slots which makes the systems less generalized for scalability with slower performance. This paper proposes an intelligent parking management system which employs deep learning to alleviate the limitations in the data driven solutions by leveraging the high performance and fast inference capability of YOLO v5 for vehicles detection instead of parking slot classification. The model was evaluated using the PKLot dataset which is a benchmark for identifying the status of the parking lot with state of the art performance achieved having an accuracy of 99.5%. To augment the performance of the algorithm in real-time, a pretrained model of YOLO v5 on MS COCO dataset was employed to detect and assign vacant parking slots and generate vehicle statistics. The performance of the proposed system was evaluated on our custom dataset with an accuracy of 96.8% achieved and a nearly real-time performance of 45 Fps which makes it more efficient, scalable, and generalized.

An Artificial Intelligence model for smart post-stroke assessment using wearable sensors

Wearable sensors provide a recording capability of the body segments’ motion. By collecting and analyzing the segments’ movements, automating the post-stroke assessment process would be feasible and reliable. Post-stroke classification of impaired or non-impaired body parts is the first step of automating the assessment process. It is recently achievable by utilizing enhancements in supervised machine learning (ML) using wearable sensors. Many classification-based approaches have numerous complexity/quality trade-offs with no optimal solution for all datasets of various structures, properties, distributions, and sizes. Ensemble machine learning provides reliable classification systems for diagnostic purposes. However, there are no criteria for identifying the baseline models in the aggregate (parallel or sequential aggregation), which significantly impacts the complexity and accuracy of the designed ensemble. In this paper, we propose an efficient post-stroke diagnosis system using the notion of multi-level ensemble learning, the Multi-Level Meta Learner (MLML) ensemble algorithm with heterogeneous or homogeneous baseline classifiers using Xsens wearable sensor collected dataset. The MLML outperforms the baseline models with comparable performance while generating a list of optimal candidate heterogeneous baseline classifiers. This list can provide only potential positive combinations, significantly improving the performance and reducing the complexity of building an ensemble. The linear acceleration and angular velocity derived from the wearable sensors are utilized. Experimental results show that the MLML has enhanced the accuracy of classifying the impaired hands from non-impaired hands with an improvement of up to 95.64% and 91.2% for boosting and Bagging learning, respectively.

Proximal Learning for Individualized Treatment Regimes Under Unmeasured Confounding

Data-driven individualized decision making has recently received increasing research interest. However, most existing methods rely on the assumption of no unmeasured confounding, which cannot be ensured in practice especially in observational studies. Motivated by the recently proposed proximal causal inference, we develop several proximal learning methods to estimate optimal individualized treatment regimes (ITRs) in the presence of unmeasured confounding. Explicitly, in terms of two types of proxy variables, we are able to establish several identification results for different classes of ITRs respectively, exhibiting the tradeoff between the risk of making untestable assumptions and the potential improvement of the value function in decision making. Based on these identification results, we propose several classification-based approaches to finding a variety of restricted in-class optimal ITRs and establish their theoretical properties. The appealing numerical performance of our proposed methods is demonstrated via extensive simulation experiments and a real data application. Supplementary materials for this article are available online.

Combining machine learning with meta-analysis for predicting cytotoxicity of micro- and nanoplastics

A growing number of articles have been reported to investigate health consequences of micro- and nanoplastics (MNPs) exposure, but no coherent images have been achieved due to the incomparability between studies, the complexity and heterogeneity of the existing toxicity data of MNPs. Herein, this work developed a predictive modelling framework for the cytotoxicity of MNPs using classification-based machine learning approaches. A literature search led to 1824 sample points represented by 9 features describing physicochemical properties of MNPs, the cell-related attribute and experimental factors. The decision tree ensemble classifier built from all features (DTE1) showed strong predictive ability with an accuracy of 0.95, recall and precision of both 0.86. Feature selection was subsequently performed to identify the key ingredients that dominated the toxic properties of MNPs. A simplified classifier developed from 6 influencing features demonstrated a comparable model performance to DTE1. This result can help direct future studies toward better experimental design and report, facilitating the understanding of the pressing MNP-related health issues. With continuous integration of more representative research data, the developed model can be widely applicable to a spectrum of MNP cytotoxicity settings.

Deep learning models for predicting the survival of patients with chondrosarcoma based on a surveillance, epidemiology, and end results analysis.

BackgroundAccurate prediction of prognosis is critical for therapeutic decisions in chondrosarcoma patients. Several prognostic models have been created utilizing multivariate Cox regression or binary classification-based machine learning approaches to predict the 3- and 5-year survival of patients with chondrosarcoma, but few studies have investigated the results of combining deep learning with time-to-event prediction. Compared with simplifying the prediction as a binary classification problem, modeling the probability of an event as a function of time by combining it with deep learning can provide better accuracy and flexibility.Materials and methodsPatients with the diagnosis of chondrosarcoma between 2000 and 2018 were extracted from the Surveillance, Epidemiology, and End Results (SEER) registry. Three algorithms—two based on neural networks (DeepSurv, neural multi-task logistic regression [NMTLR]) and one on ensemble learning (random survival forest [RSF])—were selected for training. Meanwhile, a multivariate Cox proportional hazards (CoxPH) model was also constructed for comparison. The dataset was randomly divided into training and testing datasets at a ratio of 7:3. Hyperparameter tuning was conducted through a 1000-repeated random search with 5-fold cross-validation on the training dataset. The model performance was assessed using the concordance index (C-index), Brier score, and Integrated Brier Score (IBS). The accuracy of predicting 1-, 3-, 5- and 10-year survival was evaluated using receiver operating characteristic curves (ROC), calibration curves, and the area under the ROC curves (AUC).ResultsA total of 3145 patients were finally enrolled in our study. The mean age at diagnosis was 52 ± 18 years, 1662 of the 3145 patients were male (53%), and mean survival time was 83 ± 67 months. Two deep learning models outperformed the RSF and classical CoxPH models, with the C-index on test datasets achieving values of 0.832 (DeepSurv) and 0.821 (NMTLR). The DeepSurv model produced better accuracy and calibrated survival estimates in predicting 1-, 3- 5- and 10-year survival (AUC:0.895-0.937). We deployed the DeepSurv model as a web application for use in clinical practice; it can be accessed through https://share.streamlit.io/whuh-ml/chondrosarcoma/Predict/app.py.ConclusionsTime-to-event prediction models based on deep learning algorithms are successful in predicting chondrosarcoma prognosis, with DeepSurv producing the best discriminative performance and calibration.

Classification-Based Approaches to Myopia Control in a Taiwanese Cohort.

PurposeMyopia is a disorder of growing prevalence in school-aged children worldwide, especially in Asia. Although low-dose atropine is recognized as an effective treatment to slow myopia progression, different studies have reported varying efficacies of treatment, and the optimal concentration of low-dose atropine remains an open question.MethodsA two-stage approach was conducted in this study. First, an observational study was conducted to plot the axial length growth curve for Taiwanese children. Second, an interventional 2-year study was performed in which different concentrations of low-dose atropine were applied based upon the risk-level status from the first stage.ResultsA total of 4,091 subjects, consisting of 2,105 boys (51.5%) and 1,986 girls (48.5%), were enrolled in the first stage to plot the axial growth curve for Taiwanese children aged between 3 and 16 years. The percentage of children with myopia increased from 2.3% in 4-year-olds to 88.0% in 16-year-olds. At the second stage, a total of 886 subjects [307 (34.65%) at low risk, 358 (40.41%) at moderate risk and 221 (24.94%) at high risk] were enrolled to receive low-dose atropine based upon the risk level (0.02, 0.03, and 0.05%, respectively). With this approach, the mean annual myopia progression was −0.33, −0.57, and −0.82 D in the low-risk, moderate-risk and high-risk groups, respectively. Applying annual myopic progression < -1.0 D as a criterion for responder, the responder rates were 95.77, 83.52, and 70.59% in the low-risk, moderate-risk, and high-risk groups, respectively.ConclusionsWe proposed a classification-based approach involving different concentrations of low-dose atropine based upon an individual's risk-level status. With this approach, myopic progression can be effectively controlled in patients without exposure to atropine side effects due to exposure to a higher dose than actually needed.

A longitudinal study of topic classification on Twitter.

Twitter represents a massively distributed information source over topics ranging from social and political events to entertainment and sports news. While recent work has suggested this content can be narrowed down to the personalized interests of individual users by training topic filters using standard classifiers, there remain many open questions about the efficacy of such classification-based filtering approaches. For example, over a year or more after training, how well do such classifiers generalize to future novel topical content, and are such results stable across a range of topics? In addition, how robust is a topic classifier over the time horizon, e.g., can a model trained in 1 year be used for making predictions in the subsequent year? Furthermore, what features, feature classes, and feature attributes are most critical for long-term classifier performance? To answer these questions, we collected a corpus of over 800 million English Tweets via the Twitter streaming API during 2013 and 2014 and learned topic classifiers for 10 diverse themes ranging from social issues to celebrity deaths to the “Iran nuclear deal”. The results of this long-term study of topic classifier performance provide a number of important insights, among them that: (i) such classifiers can indeed generalize to novel topical content with high precision over a year or more after training though performance degrades with time, (ii) the classes of hashtags and simple terms contain the most informative feature instances, (iii) removing tweets containing training hashtags from the validation set allows better generalization, and (iv) the simple volume of tweets by a user correlates more with their informativeness than their follower or friend count. In summary, this work provides a long-term study of topic classifiers on Twitter that further justifies classification-based topical filtering approaches while providing detailed insight into the feature properties most critical for topic classifier performance.

Assessing Representation Learning and Clustering Algorithms for Computer-Assisted Image Annotation-Simulating and Benchmarking MorphoCluster.

Image annotation is a time-consuming and costly task. Previously, we published MorphoCluster as a novel image annotation tool to address problems of conventional, classifier-based image annotation approaches: their limited efficiency, training set bias and lack of novelty detection. MorphoCluster uses clustering and similarity search to enable efficient, computer-assisted image annotation. In this work, we provide a deeper analysis of this approach. We simulate the actions of a MorphoCluster user to avoid extensive manual annotation runs. This simulation is used to test supervised, unsupervised and transfer representation learning approaches. Furthermore, shrunken k-means and partially labeled k-means, two new clustering algorithms that are tailored specifically for the MorphoCluster approach, are compared to the previously used HDBSCAN*. We find that labeled training data improve the image representations, that unsupervised learning beats transfer learning and that all three clustering algorithms are viable options, depending on whether completeness, efficiency or runtime is the priority. The simulation results support our earlier finding that MorphoCluster is very efficient and precise. Within the simulation, more than five objects per simulated click are being annotated with 95% precision.

VANT-GAN: Adversarial Learning for Discrepancy-Based Visual Attribution in Medical Imaging

• Visual attribution (VA) method is proposed for medical images. • VA is posed as discrepancy maps between an abnormal and normal counter images • Generative adversarial networks are used for abnormal-to-normal image generation. • Experiments are conducted on syntactic and Alzheimer’s disease datasets. • Proposed method outperformed baseline and related methods Visual attribution (VA) in relation to medical images is an essential aspect of modern automation-assisted diagnosis. Since it is generally not straightforward to obtain pixel-level ground-truth labelling of medical images, classification-based interpretation approaches have become the de facto standard for automated diagnosis, in which the ability of classifiers to make categorical predictions based on class-salient regions is harnessed within the learning algorithm. Such regions, however, typically constitute only a small subset of the full range of features of potential medical interest. They may hence not be useful for VA of medical images where capturing all of the disease evidence is a critical requirement. This hence motivates the proposal of a novel strategy for visual attribution that is not reliant on image classification. We instead obtain normal counterparts of abnormal images and find discrepancy maps between the two. To perform the abnormal-to-normal mapping in unsupervised way, we employ a Cycle-Consistency Generative Adversarial Network , thereby formulating visual attribution in terms of a discrepancy map that, when subtracted from the abnormal image, makes it indistinguishable from the counterpart normal image. Experiments are performed on three datasets including a synthetic, Alzheimer’s disease Neuro imaging Initiative and, BraTS dataset. We outperform baseline and related methods in both experiments.

A fragment-based structural analysis of MMP-2 inhibitors in search of meaningful structural fragments

The matrix metalloproteinase family of Zn2+-dependent metalloenzymes are a group of proteases that possess the ability to degrade the extracellular matrix (ECM) and thus are involved in different biological processes. Being one of the prime members of this family, the matrix metalloproteinase-2 (MMP-2) is associated with several pathophysiological conditions including cancer, tumor progression, invasion, and metastasis. Due to its contribution toward human pathophysiology, MMP-2 is in the limelight of anticancer drug development research for more than three decades. Our study deals with a combination of classification-based modeling approaches with fragment-based data mining techniques on a large and diverse set of MMP-2 inhibitors. This has enabled the identification of important structural fragments/features of MMP-2 inhibitors essential for their potency and efficacy. Significantly, molecular properties such as AlogP, MW, hydrogen bond features, and the molecular fractional polar surface area can be utilized to discriminate the potent MMP-2 inhibitors from the lesser active/inactive ones. From this extensive study, it can be concluded that the hydroxamate and carboxylic acid groups as suitable zinc-binding groups (ZBGs) for potent MMP-2 inhibition. This study also concludes the importance of solvent accessibility, hydrogen bond donor and acceptor groups as well as the bulky aromatic groups for better interactions inside the MMP-2 active site. Meanwhile, the similarities of this current study with our previous observations validate the outcomes of this current study. Hence, these findings may provide useful information toward the development of effective MMP-2 inhibitors in the future.

An explainable machine learning framework for lung cancer hospital length of stay prediction

This work introduces a predictive Length of Stay (LOS) framework for lung cancer patients using machine learning (ML) models. The framework proposed to deal with imbalanced datasets for classification-based approaches using electronic healthcare records (EHR). We have utilized supervised ML methods to predict lung cancer inpatients LOS during ICU hospitalization using the MIMIC-III dataset. Random Forest (RF) Model outperformed other models and achieved predicted results during the three framework phases. With clinical significance features selection, over-sampling methods (SMOTE and ADASYN) achieved the highest AUC results (98% with CI 95%: 95.3–100%, and 100% respectively). The combination of Over-sampling and under-sampling achieved the second-highest AUC results (98%, with CI 95%: 95.3–100%, and 97%, CI 95%: 93.7–100% SMOTE-Tomek, and SMOTE-ENN respectively). Under-sampling methods reported the least important AUC results (50%, with CI 95%: 40.2–59.8%) for both (ENN and Tomek- Links). Using ML explainable technique called SHAP, we explained the outcome of the predictive model (RF) with SMOTE class balancing technique to understand the most significant clinical features that contributed to predicting lung cancer LOS with the RF model. Our promising framework allows us to employ ML techniques in-hospital clinical information systems to predict lung cancer admissions into ICU.

An unsupervised method for social network spammer detection based on user information interests

Online Social Networks (OSNs) are a popular platform for communication and collaboration. Spammers are highly active in OSNs. Uncovering spammers has become one of the most challenging problems in OSNs. Classification-based supervised approaches are the most commonly used method for detecting spammers. Classification-based systems suffer from limitations of “data labelling”, “spam drift”, “imbalanced datasets” and “data fabrication”. These limitations effect the accuracy of a classifier’s detection. An unsupervised approach does not require labelled datasets. We aim to address the limitation of data labelling and spam drifting through an unsupervised approach.We present a pure unsupervised approach for spammer detection based on the peer acceptance of a user in a social network to distinguish spammers from genuine users. The peer acceptance of a user to another user is calculated based on common shared interests over multiple shared topics between the two users. The main contribution of this paper is the introduction of a pure unsupervised spammer detection approach based on users’ peer acceptance. Our approach does not require labelled training datasets. While it does not better the accuracy of supervised classification-based approaches, our approach has become a successful alternative for traditional classifiers for spam detection by achieving an accuracy of 96.9%.

Weighted Deep Neural Network Ensemble Approach for Multi-Domain Sentiment Analysis

The sentiment analysis is a subtask of text classification that is known as a domain dependent problem. In order to obtain an accurate classifier for a particular domain, a large labelled dataset is needed. To tackle the challenge of data scarcity in some domains, in the area of multi-domain problems, the classifier is trained on a set of labelled data from some domains and then it is applied to the target domains. In addition, another important issue in classification-based approaches in order to reach the better performance is that the nature of train and test data should be similar. So, a model trained by data from a specific domain, leads to poor results when it comes to another domains. This paper proposes three Weighted(deep)Neural Networks Ensemble approaches for multi-domain sentiment classification to address the mentioned issues, by training individual deep learning models (including CNN, LSTM and Bi-GRUCapsule) on specific domains. Using a weighted score of DBD and the initial polarity of the sample test data on each domain, a new aggregated score of final polarity is obtained. The DRANZIERA protocol is used for evaluation of the proposed models. The results have shown more than 0.03 improvements in average accuracy in comparison to the other state-of-the-art approaches.

Siamese networks for large-scale author identification

Authorship attribution is the process of identifying the author of a text. Approaches to tackling it have been conventionally divided into classification-based ones, which work well for small numbers of candidate authors, and similarity-based methods, which are applicable for larger numbers of authors or for authors beyond the training set; these existing similarity-based methods have only embodied static notions of similarity. Deep learning methods, which blur the boundaries between classification-based and similarity-based approaches, are promising in terms of ability to learn a notion of similarity, but have previously only been used in a conventional small-closed-class classification setup.Siamese networks have been used to develop learned notions of similarity in one-shot image tasks, and also for tasks of mostly semantic relatedness in NLP. We examine their application to the stylistic task of authorship attribution on datasets with large numbers of authors, looking at multiple energy functions and neural network architectures, and show that they can substantially outperform previous approaches.

Cytotoxicity of phytosynthesized silver nanoparticles: A meta-analysis by machine learning algorithms

The rapid development and increasing use of silver nanoparticles (AgNPs) synthesized by “green” methods such as by plants for biomedical, textile and healthcare applications has raised questions regarding their potential impacts on human health. This study presents a meta-analysis of the cytotoxicity data of phytosynthesized AgNPs with heterogenous features from literature using two classification-based machine learning approaches, decision tree (DT) and random forest (RF). The inclusion of plant family as a biosynthesis-related feature clearly improved the accuracy and generalization performance of DT and RF models, revealing the potential impact of biosynthesizing parameters on the cytotoxic effects of phytosynthesized AgNPs. A measure of the mean decrease Gini in the RF modeling identified that exposure regime (including time and dose), plant family, and cell type were the most important predictors for the cell viability outcomes of green AgNPs. Further, the potential effects of major variables (cell assays, intrinsic nanoparticle properties, and reaction parameters used in biosynthesis) on AgNPs-mediated cytotoxicity and model performance were discussed to provide a basis for future work. Thus, this meta-analysis of published data by machine learning algorithms provides guidance and prediction to key variables affecting AgNP-mediated cytotoxicity, which may help direct future studies toward better experimental design as well as the virtual design or optimization of green AgNPs for specific applications.

An approach for discovering firm-specific technology opportunities: Application of link prediction to F-term networks

Technology opportunity discovery (TOD) based on firm's technology portfolio is categorized into text mining-based and patent classification-based approaches. Despite their apparent benefits, the former has reproducibility issues due to the experts’ subjectivity, whereas the latter lacks consideration of technical attributes that constitute individual technologies. The F-term, a multi-dimensional subject indexing system, provides patent classification codes representing technical attributes and structures that vary according to subject technology. The present study proposes a TOD model employing a link prediction analysis of F-terms. The proposed model based on F-terms comprises four steps: 1) constructing a universal F-term network using F-term co-occurrences; 2) generating a firm-centered F-term network highlighting a target firm's technology portfolio; 3) applying a proposed link prediction index to identify opportunity F-terms; and 4) assessing these opportunities in terms of technical attributes using a visual map with technology impact and heterogeneity indices. A case study is conducted on a Japanese firm to demonstrate the function and validity of this model, which aims to assist firms to identify technology opportunities with high practicality considering both their technology portfolios and the entire technology ecology. Moreover, this study represents a contribution to an early attempt to apply large-scale F-terms to the quantitative TOD.