Abstract

Although artificial intelligence can contribute to decision-making processes, many industry players lag behind pioneering companies in utilizing artificial intelligence-driven technologies, which is a significant problem. Explainable artificial intelligence can be a viable solution to mitigate this problem. This paper proposes a research model to address how explainable artificial intelligence can impact decision-making processes. Using an experimental design, empirical data is collected to test the research model. This paper is one of the pioneer papers providing empirical evidence about the impact of explainable artificial intelligence on supply chain decision-making processes. We propose a serial mediation path, which includes transparency and agile decision-making. Findings reveal that explainable artificial intelligence enhances transparency, thereby significantly contributing to agile decision-making for improving cyber resilience during supply chain cyberattacks. Moreover, we conduct a post hoc analysis using text analysis to explore the themes present in tweets discussing explainable artificial intelligence in decision support systems. The results indicate a predominantly positive attitude towards explainable artificial intelligence within these systems. Furthermore, the text analysis reveals two main themes that emphasize the importance of transparency, explainability, and interpretability in explainable artificial intelligence.

Introduction

Automotive industry leaders, such as Tesla, have made substantial investments in artificial intelligence (AI) to expedite the introduction of self-driving vehicles to the market, enhancing their competitive capabilities. The integration of AI in supply chain operations has played a crucial role in enabling Tesla to optimize its operational costs [64] while simultaneously facilitating the establishment of a Gigafactory in China [78]. We are witnessing a rapid digital transformation driven by the integration of AI in supply chain management. The COVID-19 pandemic forced companies and organizations to expedite the digitalization of their operations [6]. To enhance their competitive edge, prominent companies, including Amazon, Walmart, Alibaba, Siemens, and Toyota have embraced AI-based technologies to automate and digitalize their operations and supply chain activities [1,32]. Digital transformation also introduces new possibilities for potential cyberattacks. However, using AI-based technologies for decision-making during cyberattacks (e.g., American Express monitoring; [75]) offers significant advantages that outweigh the potential losses incurred.

Although AI-based technologies can contribute to decision-making processes in operations and supply chain management, many industry players are lagging behind pioneering companies in utilizing AI-driven technologies, which is a main problem. Therefore, the adoption of AI-based technologies in decision-making processes, particularly during sensitive situations such as cyberattacks, may encounter potential barriers that delay their usage. Due to its vital advantages, AI has received much attention from decision-makers to address resilient case studies [45] and other sensitive problems such as healthcare [69]. The lack of explanations of the underlying AI processes leads to the rejection of AI in decision support systems [60]. Leveraging AI-powered decision-making platforms can significantly facilitate and expedite the decision-making process, which results in improved overall performance. For instance, the Colonial Pipeline, a U.S. oil supplier, faced a cyberattack and, after a week of deliberation, opted to pay around $4.4 million to solve the issue [21]. A quick decision on the first day through agile decision-making could have saved them money and enabled the uninterrupted continuation of their operations with stockholders.

Results

We employed regression analysis to test the hypothesized relationship between research model variables. We conducted required tests relating to the experimental design along with method assumptions and bias checks.

4.1. Design check
To check the design, the scenario test, the realism test, the validity test, and the manipulation test were performed. We conducted the scenario check to measure the participants’ learning process regarding the provided information. The results showed that participants’ learning was strong enough to support the scenario design (in the first scenario, the Mean = 5.3 out of 7, with a Standard Deviation = 0.99, and in the second scenario, the Mean = 5.5 out of 7, with a Standard Deviation = 0.85). The realism check was conducted through a face validity process and tested using the quantitative approach proposed by Thomas et al. [65]. Participants supported the realism check by agreeing that the scenarios were close enough to reality (Mean = 4.08 out of 5; Standard Deviation = 0.44). Similarly, the validity check supported the scenario design (Mean = 4.07 out of 5; Standard Deviation = 0.43). The manipulation test supported the manipulation by showing a significant difference between the two manipulated levels (p-value <0.001, the first level, the low level of XAI: Mean = 2.68 out of 7; Standard Deviation = 0.12; and the second level, the high level of XAI: Mean = 5.9 out of 7; Standard Deviation = 0.10).

Abstract

     The mission of the Institute of Industrial and Systems Engineers (IISE) is to serve those who solve complex and critical problems of the world. Notably, the research–practice gap in Operations Management (OM) marginalizes the value and relevance of the IISE. To maintain and enhance the impact of the IISE, we identify major bottlenecks that limit the industrial installation of OM research outcomes. Ranked by the relative importance, the three bottlenecks are verifying the performance improvement, building trust with practitioners, and balancing model accuracy and simplicity, respectively, in the stages of value verification, implementation and development. We propose potential research opportunities and illustrate the challenges and opportunities using real case studies from three Fortune Global 500 companies. In particular, we emphasize the role of data-driven decision methods in dealing with the three bottlenecks.

Introduction

     It is crucial for the Institute of Industrial and Systems Engineers (IISE) to guide Operations Management (OM) practitioners toward achieving OM practice excellence. The mission of the IISE is to serve OM practitioners who solve complex and critical problems encountered in the world, and often need to deal with complicated and integrated systems that suffer from customer dissatisfaction, excess inventory costs, and low product quality. Therefore, it is imperative for the IISE to capture the full potential of successful OM practice.

     However, the academia that studies OM has generally been following, rather than leading, business practices (Simchi-Levi, 2014). The inability to lead OM practices is often due to numerous failed implementations of OM models. After many years of OM projects with industry collaborators, we cannot emphasize enough how important a successful implementation is for a research project. Further, we are astonished by the difficulties associated with ensuring an OM implementation with precise and unified standards across different departments. Companies have been appropriately cautious about new OM models because they cannot fully reap its benefits (Ibanez et al., 2018; Sun et al., 2021).

Conclusions

     The future is promising as more OM research tends to bring the theoretical studies close to practical issues. The authors have extracted the data of all the 300 OM-related projects funded by the National Natural Science Foundation of China (NSFC) from 2016 to 2021. Table 1 categorizes the subjects of all the projects by new business model, new technology & Supply Chain Management (SCM) finance, product & behavior, green & health-care & agriculture, and traditional SCM. It is surprising to see that almost all the subjects are driven by real practices, and few project proposals even involve preliminary plans for industrial implementations.

     The academic community has always emphasized novel and mathematically provable methods and reproducible results, yet none of which are required for complex realworld issues (Sodhi and Tang, 2008). To thrive, the ISE research community should focus on overcoming the challenges that hinder successful practical applications and embrace cooperation with data-driven decision methods so that our research community will stay relevant and impactful. Combined with real case studies, we list and analyze the key challenges. Ultimately, these challenges can stimulate research opportunities such as user-centric algorithms and explainability. We hope that this article will encourage more ISE researchers to embark on research that can help close the gap when it comes to implementing new OM research ideas into practice.

Abstract

     In parallel with the development of information and network technology, large amounts of data are being generated by the Internet, and data-driven methodologies are now often being used in decision-making. Recent studies have investigated personalized individual semantics (PIS) in various decision-making contexts to model a fact that words mean different things to different people. However, few studies have investigated PIS in the context of multi-attribute decision-making (MADM). In MADM, in addition to multi-attribute linguistic information, pre-existing classification of the alternatives is always present, which have not been considered in prior research. Most previous studies have simply demonstrated the feasibility of PIS methods with numerical examples using small-scale models, and not with realistic datasets. Therefore, in this study, we propose a data-driven learning model to analyze the PIS of decision makers to support a multi-attribute decision-making model that considers pre-existing classification of the alternatives. Specifically, we first propose a PIS multi-attribute learning function to define a general computation form for comprehensive evaluation of the value of alternatives. Then, considering this pre-existing classification of the alternatives, a PIS learning model is constructed by analyzing the relations between calculated values of alternatives and corresponding class assignments to obtain personalized numerical scales of linguistic terms for a decision maker. Finally, we present a case study based on two datasets and a comparison with other methods to justify the feasibility of the proposed model.

Introduction

     Multi-attribute decision-making (MADM) involves an individual or a group of decision makers selecting an option by evaluating a set of alternatives according to multiple attributes [1,2,3,4]. In realistic MADM problems, decision makers often prefer to use language and linguistic terms to express their preferences for evaluating the objects or alternatives, and such information is conventionally included in multi-attribute linguistic decision matrices to represent their preferences [5].

     This means there is a need for computing with words (CWW) [6,7] when dealing with linguistic preferences in decision making. Obviously, words mean different things to different people [8,9], and type-2 fuzzy sets [8] are commonly used to deal with this issue in CWW. The CWW model of type-2 fuzzy sets is a useful tool to deal with multiple meanings, but this model does not represent specific semantics for individuals. For example, if a family with three members wants to buy a car, perhaps the members of the family all think the car is “good”. However, the numerical decision-making meaning of the word “good” might be accurately represented as 0.9 for one member, and 0.7 for the other two members. This reflects PIS among the three members. Porro et al. [10] proposed the concept of perceptual maps to represent the differences between the decision makers’ semantics of linguistic terms in MADM. Recently, Li et al.

Conclusions

     In this study, considering the fact that words mean different things to different people, we propose a PIS learning method for DD-LMADM based on multi-attribute linguistic information and information on the classification of alternatives. The proposed model starts with a PIS multi-attribute learning function to provide a general computation form for evaluating the comprehensive values of alternatives in LMADM models; based on this function, we propose a data-driven PIS learning model that considers the distinct linguistic term sets and the unknown weights associated with the criteria to calculate the PNS of linguistic terms for decision makers.

     The practical examples with car evaluation datasets and house evaluation datasets illustrate the applicability of the proposed data-driven PIS learning approach. Furthermore, we compare our approach with the existing CWW methods based on a triangular membership function and a 2-tuple linguistic model in terms of the inconsistency between the calculated preference values of alternatives and their class assignment.

Abstract

     This study integrated multiple objective decision-making approaches including data envelopment analysis (DEA), rough set theory and TOPSIS method for exploring the competitiveness of pharmaceutical multinational enterprises. Firstly, this study applied an advanced two-stage network DEA to measure the R&D efficiency and business performance of pharmaceutical multinational enterprises (PMNEs) listed in the Forbes Global 2000 and ranked these PMNEs by using rough set theory, DEA and TOPSIS method. In addition, differences in environmental, social, and governance (ESG) performance across three continents were investigated. Findings show that North America is significantly the best (worst) region in terms of business performance (R&D efficiency) while Europe is significantly the best (worst) region in terms of R&D efficiency (business performance). Alfresa Holdings Corporation, a PMNE from Japan, is the only PMNE that is efficient from both an R&D and business performance perspective. European PMNEs have significantly better environmental and social performance than other regions, however, they have the worst governance performance. Overall, this study provides insights to managers and investors into the application of various methods for accurately measuring performance and ranking multinational enterprises.

Introduction

     The pharmaceutical industry is one of the most important and competitive industries in the world, it not only has a critical role in the health maintenance of individuals but also in economic growth (Sharma & Modgil, 2019). Pharmaceutical firms have spent a substantial proportion of their time and investment in research and development (R&D) (Rao, 2020). In 2018, R&D expenditure in the pharmaceutical industry was 179 billion U.S. dollars globally (Mikulic, 2020). A correct grasp of the R&D efficiency of the industry is required to improve resource allocation and thereby prevent either excess or shortage of resource inputs (Liu & Lyu, 2020). According to Lu et al. (2019), evaluating multiple indicators simultaneously is useful for appropriate decision-making. An overall efficiency can be decomposed into two stages of efficiencies (Kao & Hwang, 2008), suggesting that we should also evaluate the sub-processes of a firm (Kao & Hwang, 2011). That is, managers should evaluate their firm performance not only from a multidimensional perspective but also from those of its internal processes. As mentioned in the research of Sharma and Modgil (2019), business performance additionally depends on cost reduction of manufacturing operations and improving the firm’s innovative ability. Recently, Hsieh et al. (2020) have applied two-stage data envelopment analysis (DEA) to investigate organizations’ innovation and business performance by investigating the network structures of decision-making units (DMUs).

Conclusion

     This study measures the firm performance of 41 PMNEs from a two-stage perspective, namely R&D efficiency and business performance, by integrating a two-stage network DEA and SOCP technique. Moreover, this study integrates the DEA approach and rough set theory to rank PMNEs. The findings of this research can serve as a reference for managers of PMNEs in their decision-making process. Besides, we examine differences in the ESG scores of the PMNEs among different geographic regions (America, Asia, and Europe). This acts as a guide for investors, managers, and other stakeholders when attempting to move their ESG initiatives further.

     R&D performances in the operation of the observed PMNEs are not attracting enough attention. Alfresa Holdings Corporation is the only efficient PMNE for both R&D efficiency and business performance. North America is significantly the best (worst) region in terms of business performance (R&D efficiency) while Europe is significantly the best (worst) region in terms of R&D efficiency (business performance). From the perspective of ESG performance, European PMNEs significantly have better environmental and social performance than other regions whereas they are the worst concerning governance performance.

Abstract

     Concrete is a primary material in infrastructure projects and is a significant contributor to global climate emissions. However, there is a lack of readily available cement and ready-mix concrete inventory data for evaluating the environmental performance of the industries. This study describes the development of cradle-to-gate inventories for U.S. ready-mix concrete and gate-to-gate inventories for portland cement production technologies. These life-cycle inventories provide baselines for the environmental releases associated with concrete that is used for major infrastructure projects. The inventories are incorporated into the publicly available Greenhouse gases, Regulated Emissions, and Energy use in Technologies (GREET) model. The life-cycle inventories are created using facility-level environmental release data from U.S. datasets normalized to activity levels which are based on production capacity and utilization data provided by Portland Cement Association (PCA) and the U.S. Geological Survey Minerals Yearbook. Unit processes for limestone quarrying, sand and gravel quarrying, and wet-mix concrete batch plants are developed on the basis of national total point-source environmental releases and production statistics, coupled with corresponding flows associated with off-road fuel consumption and other non-point-source emissions. Midpoint impact assessment results are normalized to provide insight into their relative significance in the context of U.S.

Introduction

     Concrete is a primary structural material central to the construction projects throughout the world. It is the second most-consumed material in the world, after water (Miller et al., 2018b). Cement is the binder that holds concrete aggregates together, enabling structural performance. Global cement production of about 4.1 billion tonnes year after year since 2013 (IEA, 2021). Of this, 102 million tonnes, or 2.5%, were consumed in the U.S. Concrete can be transported, pumped, and poured to fill frames in a wide range of forms, and once cured, it is rigid and durable, making it ideal for structural elements, especially distributed transportation infrastructure. As a result, concrete is used in a wide range of construction projects. The Portland Cement Association (PCA) reports that in 2015, 30% of the cement use in the U.S. was for transportation infrastructure, 28% for residential buildings, 18% for nonresidential buildings, 13% for public utilities and water/wastewater systems, and 10% for other applications (PCA, 2016).

Conclusions

     This study provides transparent emissions data for infrastructure modeling within the GREET framework, as well as robust inventories for openLCA modeling, available for free download on the Federal LCA Commons. The four primary kiln technologies employed to produce cement are all available, in addition to the U.S. national average blend, enabling specification of blends that may be more regionally appropriate for specific LCA studies. The ready-mix concrete inventory is appropriate for commonly available truck-delivered concrete and is also adaptable for specified cement blends, including adjustments for modeling batch plants that may be set up for larger infrastructure projects. These data enable transparent assessments of concrete applications with complete inventories across a full suite of life-cycle impact categories. Researchers and decision-makers can use GREET or the data on the Federal Commons to assess the impacts associated with major concrete infrastructure projects.

Abstract

References

Acknowledgements

Author information

Ethics declarations

Additional information

Rights and permissions

About this article

Abstract

The purpose is to make full use of data mining and machine learning technology under big data to improve the ability of trade financial enterprises to cope with the risk of excessive financialization. In view of the above needs, based on previous studies, genetic algorithm (GA), neural network and principal component analysis (PCA) methods are used to collect and process the data, and build a risk assessment model of excessive financialization of financial enterprises. The performance of the model is analyzed through the data of specific cases. The results suggest that the data mining technology based on back propagation neural network (BPNN) can optimize the input variables and effectively extract the hidden information from the data. The specific examples show that most of the current enterprises do not have greater financial risk. However, most of the financial enterprise indexes show that the actual enterprise assets are gradually financialized. The total accuracy rate of financial risk assessment model based on deep belief network (DBN) is over 91%, and the accuracy of the model can reach 80% even if the sample size is small. Therefore, the financial risk assessment model proposed can effectively analyze the relevant financial data, and provide reference for the financial decision-making research of financial enterprises.

1 Introduction

With the continuous development of the world economy in recent years, more and more enterprises are active in the fnancial market, which promotes the sustainable development of the national economy. However, the following problems are the fracture of many enterprise capital chains due to the poor risk awareness, fnancial market turbulence, commercial fraud or mismanagement. Small enterprises are on the verge of bankruptcy because of their poor ability to resist risks, which is mainly because more enterprises do not invest in the real economy but pay more attention to virtual fnancial markets (Anginer et al., 2018). With the advent of big data era, although the risk control model of enterprises plays a certain role in credit management, due to the lack of deep understanding of data, the corresponding model accuracy is low, and it is unable to evaluate the borrower comprehensively (Florio & Leoni, 2017). Due to the more in-depth research on big data tools by many scholars, more and more methods are applied to the fnancial feld (Trelewicz, 2017). Traditional risk control model has problems such as single dimension and limited assessment ability. However, using data mining method can analyze the data of fnancial industry in depth, involving many dimensions and more comprehensive assessment (Moradi & Mokhatab Rafei, 2019). With the establishment of distributed database and the gradual maturity of various data platform architectures, more data are stored and calculated, and the dimensions of data analysis are also wider (Medvedev et  al., 2017). Especially with the construction of Hadoop cloud computing platform, the distributed computing ability of big data has been greatly improved, and the data foundation and computing ability are more efcient. Compared with the traditional risk control model, the model based on big data has great changes in the amount of information and accuracy (Ding et al., 2019; Ivanov et al., 2019). Big data can improve the credit system and help trade enterprises reduce credit risk (Saura et  al., 2019; Urbinati et  al., 2019). Data mining can use the network data to improve the scoring model and optimize the approval process through the integration of algorithms, so as to form a sustainable closed-loop management mode (Aldridge, 2019). Therefore, the use of data mining and its application in fnancial risk assessment has become one of the hot spots in this feld.

With the continuous development of Internet technology and electronic information technology, big data technology and cloud computing technology also rise and develop, and have a positive impact on people’s lives. Data mining system can deeply mine the value information contained in big data, so as to assist users to solve practical problems and help users to make correct decisions and judgments. This paper presents an energy analysis of data mining algorithm based on cloud platform for Internet of things (IoT). First of all, an improved Apriori algorithm is proposed, which is based on Boolean matrix and sorting index rules. Then Boolean matrix is obtained after scanning the dataset and the Boolean matrix is preprocessed to delete the useless transactions and the item set, which are combined with sorting index to produce other item sets, effectively improving the efficiency of frequent item mining, which effectively reduce the memory usage. Secondly, the density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm needs human intervention in the global parameter selection, and the process of regional query is complex and the query is easy to lose objects. An improved parameter adaptive and regional query density clustering algorithm is proposed, which can effectively delete the redundant data in the high-level complex data space on the premise of retaining the internal nonlinear structure of the IoT data. The efficiency of clustering is also improved accordingly. Finally, the simulation based on cloud platform verifies the effectiveness and superiority of the algorithm.

Introduction

With the continuous development of Internet and the intelligent terminals, the amount of data has become more and more, and the types of data have become extremely rich, which represents that human beings have been big data age [1]. In recent years, data analysis of the Internet of things (IoT) has attracted more and more [2]. Through effective data analysis methods, the characteristics and rules of the Internet of things can be mined to reduce the loss in each life cycle of the Internet of things [3]. At present, the traditional Internet of things data analysis method needs to establish a high-precision mathematical model of the IoT data system, and the system is limited to the linear and time invariant system [4]. At the same time, it has high requirements for knowledge in the professional field and low precision shortcomings, so it has great limitations in application. At present, the collection mechanism of Internet of things data in many areas is becoming more and more perfect [5]. The Internet of things database will collect Internet of things data, which contains a lot of undiscovered massages, and the Internet of things data has many disadvantages, but it is difficult for traditional Internet of things data analysis methods to find and summarize the content of these Internet of things data [6]. Knowledge discovery, machine learning and data mining technology are just new disciplines born under this background. Because data mining technology has the ability to deal with large amount of data and find potential massages, it is a way to mine important massages in IoT data [7].

Introduction

Research contents

Research results

Conclusion

References

Abstract

The purpose of this study was to analyze the characteristics of the diagnosis and treatment of female uterine fibroids based on data mining technology and the medication rule. Clinical literature related to the treatment of uterine fibroids by Chinese National Knowledge Infrastructure (CNKI) from 2004 to 2018 were searched, and the literature meeting the requirements were selected according to the inclusion criteria and exclusion criteria. The contents of the prescription were recorded into the database, and the results of drug frequency, common drugs and core rules, etc. were obtained by the rule analysis of the software, so as to analyze the medication rules of modern doctors in treating uterine fibroids. Through research and analysis, it showed that common clinical drugs include leuprorelin (GnRH-a), danazol (androgen), gestrinone (progestin), mifepristone (progesterone receptor antagonist), and some cases of combination therapy. Based on the analysis of the frequency of drug, treatment effect and adverse drug reactions, it was found that the clinical application of mifepristone in the treatment of uterine fibroids was relatively common, with significant drug effect and mild adverse reactions, which was worthy of clinical promotion and application. Therefore, the results of this study provide a new basis for the clinical and basic research of traditional Chinese medicine in the treatment of uterine fibroids, but it still needs the verification of expert interpretation, experimental research and other methods.

Introduction

Uterine fibroids are one of the most common benign tumors in women and are derived from the monoclonal proliferation of uterine smooth muscle cells. Uterine fibroids are more common in women of childbearing age and are greatly affected by estrogen and progesterone. The incidence of uterine fibroids is related to race, with the highest incidence among African Americans, accounting for 80% [1]. In addition, the high-risk factors of fibroids also include age, early menarche, late childbirth, coffee, alcohol abuse, gene mutation, obesity and other problems [2]. There are a variety of clinical manifestations of uterine fibroids, including menorrhagia, prolonged menstruation, pain, anemia, abortion, infertility, abnormal band, bladder and rectal compression symptoms. Uterine fibroids will not only affect the fertility of patients, but also affect the quality of life and psychological status of patients. In terms of infertility, infertility caused by uterine fibroids alone accounts for less than 3% of all infertile patients [3,4]. Studies have shown that normal tissue around the uterus in patients with uterine fibroids contains significantly less estrogen and progesterone receptors than fibroids. In patients with uterine fibroids, the normal tissue surrounding the uterus contains significantly reduced estrogen and progesterone receptors than in fibroid tissue.In 2014,the state food and drug administration has approved Mifepristone (RU486), gonadotropin-releasing hormone agonist (GnRH-a), gossypol and some traditional Chinese medicines for the treatment of uterine fibroids [5,6]. Recent studies have shown that mifepristone not only directly inhibits follicular development and secretion in the ovary, but also may inhibit the secretion of follicle stimulating hormone (FSH) and luteinizing hormone (LH). Therefore, the drugs commonly used in the clinical treatment of uterine fibroids are drugs that can reduce the level of estrogen and progesterone. For example, the use of gonadotropin receptor agonists.

1- Introduction

2- Definition of AI, ML and DM

3- Application of DM methods for managing production complexity

4- Findings and conclusion

References

Abstract

Production complexity has increased considerably in recent years due to increasing customer requirements for individual products. At the same time, continuous digitization has led to the recording of extensive, granular production data. Research claims that using production data in data mining methods can lead to managing production complexity effectively. However, manufacturing companies widely do not use such data mining methods. In order to support manufacturing companies in utilizing data mining, this paper presents both a literature review on definitions of data mining, artificial intelligence and machine learning as well as a categorization of existing approaches of applying data mining to manage production complexity.

Findings and conclusion

At the interface of rising production complexity due to shifting market demands and vast amounts of production data, DM can be a valid tool to support managing complexity. Most applications of DM in production management have so far been related to quality management. There are very few applications of DM directly related to production complexity. However, other applications of DM in other fields of production management serve the purpose of managing production complexity very well. We have presented some of these applications and plan to extend the categories in future work to present a holistic framework of DM, as well as other ML and AI applications able to cover all relevant aspects of managing production complexity.

1- Introduction

2- Structural databases

3- Data mining techniques

4- Examples of data mining energy materials from structure databases

5- Suggestions and outlook

6- Conclusions

References

Abstract

New energy materials that act as clean power sources and data science are developing rapidly in the past decades and the advancement of the two research areas have significantly benefited the development of each other. At the meantime, structural information of materials have been obtained and stored in various structure databases, such as the Cambridge Structure Database (CSD) and the Inorganic Crystal Structure Database (ICSD). Researchers have developed various structure-property relationships of the energy materials, which could be applied to screen the potential suitable materials from structure databases; this has become an efficient route to explore and design new energy materials. In this article, we review recent progresses on the data mining study of new energy materials based on structure databases such as CSD and ICSD, in the context of dye-sensitized solar cells and perovskite solar cells, and also include other energy systems such as water splitting systems, lithium batteries, thermoelectric devices and gas adsorbent materials. The structure descriptors that are more fundamental in the data mining procedure employing the structure-properties relationships are focused; the structural descriptors are complementary to the quantum descriptors and are efficient in the materials design process. We believe that with the successful formulation of more advanced and case-by-case structure-property relationships of energy materials, many new energy materials could be efficiently identified with much lower cost and shorter design period via the data mining process.

Introduction

There has been global energy crisis and environmental issues in the past decades due to the overuse of fossil fuels. New energy materials should be identified and developed to provide clean energy [1–5]. For example, efficient solar cell materials should be explored to harness the clean energy from the non-exhaustible solar system to provide the electricity; materials incorporated in water-splitting systems should be screened to provide solar fuels; lithium-based materials could be exploited to be tailored to undergo charge/discharge process reversibly and store electrical energies for portable devices and vehicles; metalorganic frameworks could be discovered to capture the carbon and gas molecules effectively; thermoelectric materials and piezoelectric materials should be identified to covert the thermal and mechanical energy into the electric current. At the current stage, the discovery of these energy materials relies predominantly on the experimental serendipity and the try-and-error experimental process that are inefficient and time-consuming. Nevertheless, many evidences have shown that the discovery process of new energy materials could be greatly accelerated by the data mining process, which have already shown their excellence in pharmaceutical drug discovery, finance, medicine, and marketing [6,7]. A large amount of structure information of new materials have been determined and stored in structure databases, which are mainly prepared by the crystallization processes and solved by the X-ray diffraction techniques and other crystallographic techniques [8]. Most of these materials have their structures accurately determined within the resolution of 0.01 Å and serve as outstanding platforms for the structural analysis and more advanced structure-property analysis. Two outstanding databases of materials structures have emerged: the Cambridge Structure Database (CSD) and the Inorganic Crystal Structure Database (ICSD), which focus on the accurate crystal structures of organic and inorganic materials, respectively [9,10]. These databases formed a good foundation for the data mining process of new energy materials. These databases store structures of materials that have already been synthesized in laboratory and exhibit significant advantages to expedite the process towards achieving promising candidates [11,12].