Bayes Classification Approach Research Articles

Sentiment Analysis of e-Government Service Using the Naive Bayes Algorithm

A digital platform called the Alpukat Population Application is used to handle statistics and information regarding DKI Jakarta's population. Using the Naive Bayes Classifier (NBC) approach, sentiment analysis for applications using satellite placement. The Nave Bayes Classifier technique is utilized for sentiment analysis because of its benefits in modeling and categorizing complicated data. The user reviews and comments gathered from the Google Play Store were the source of the data utilized in this research. Feature extraction using methods like TF-IDF, sentiment labeling on data, and the development of Nave Bayes Classifier models for sentiment classification were all part of the research project. It is anticipated that the study's findings would help us better understand how users interact with the Alpukat population app. This sentiment analysis may assist app administrators and developers in identifying the positives and negatives of applications and planning updates and advancements based on user feedback. It is anticipated that the sentiment classification model created using the Naive Bayes Classifier approach would be able to classify user evaluations into positive, negative, or neutral sentiment categories with a high degree of accuracy. The creation of improved alpukat positioning apps and decision-making may both benefit from this emotive analysis.

Intrusion Detection in Internet of Things Systems: A Feature Extraction with Naive Bayes Classifier Approach

The Internet of Things (IoT) has proliferated, transitioning from modest home automation to encompass sectors like healthcare, agriculture, transportation, and manufacturing. This evolution is characterized by devices' ability to autonomously gather, disseminate, and analyze data, leading to improved real-time decision-making, predictive insights, and customized user experiences. The ubiquity of IoT, while promising, introduces significant data security concerns. The vast number of interlinked devices and diverse and often insufficient security features make them vulnerable to cyber threats, emphasizing the need for robust security mechanisms. Intrusion Detection Systems (IDS) have traditionally acted as vital guards against such threats; however, with the ever-increasing data in the IoT, traditional IDS models, such as Naive Bayes, face processing speed and accuracy challenges. This paper introduces a novel model, "FE+NB," which merges advanced Feature Extraction (FE) with the Naive Bayes (NB) classifier. Central to this model is the "Temporal-Structural Synthesis" technique tailored for IoT traffic data, focusing on data compression, temporal and structural analyses, and Feature Selection (FS) using mutual information. Consequently, the model enhances efficiency and accuracy in Intrusion Detection (ID) in complex IoT networks.

Analisis Sentimen Ulasan Pada Aplikasi E-Commerce Shopee Dengan Menggunakan Algoritma Naïve Bayes

Every year, the e-commerce industry in Indonesia grows and develops at a rapid pace. In Indonesia, many online marketplaces have sprung up, including Tokopedia, Lazada, and Shopee. People are very interested in all e-commerce companies because they are hassle-free and instant. Among the most famous is Shopee, which offers a wide range of services and also presents a ratings and reviews column. This feature allows users to express their feelings about Shopee. Based on the information gathered from previous user comments, consumers can use these ratings to identify and trust both excellent and negative recommendations of the app they want to use. Sentiment analysis results include both favorable and negative user reviews by scoring, classifying, and filtering viewpoints to help businesses or users. The author uses the naïve bayes algorithm in this research. The Naïve Bayes Classifier approach will be used in this research to perform sentiment classification. The author then uses associations between frequently discussed word terms or themes that are related to each other for the extraction and exploration process, as well as descriptive statistics. Naïve Bayes Classifier is a binary classification technique that applies Bayesian principles with a strong assumption of independence, utilizing simple statistical probabilities.

Analyzing the Impact of Body Shaming on Twitter: A Study Using Naive Bayes Classifier and Machine Learning

Twitter is one type of social media that is still widely used today. However, it happens frequently for Twitter users to post remarks that tend to criticize other Twitter users. Twitter users routinely make nasty remarks regarding body shaming, which has detrimental impacts on the victims such as diminishing self-esteem, leading to depressive illnesses, and, more seriously, raising the chance of suicide. Body shaming is the practice of criticizing someone's physical attributes, such as being slim, overweight, or having a flat nose. This study will use the Naive Bayes Classifier approach to do sentiment analysis based on the actions of body shaming on Twitter. Based on the performance testing results of Accuracy, Precision, and Recall using Machine Learning Rapid Miner with an initial dataset of 1000 body shaming tweets and a test dataset of 329 tweets, the following results were obtained: Accuracy of 80.55%, positive Precision of 100%, negative Precision of 80.43%, positive Recall of 3.03%, and negative Recall of 100%. In the preprocessing stage, tokenization resulted in a word cloud with the top 5 words being "overweight" at 51%, "body shaming" at 20%, "thin" at 11%, "people" at 10%, and "eating" at 8%.

DETERMINANTS OF NATURAL RESOURCES BASED MICROENTERPRISES PERFORMANCE IN INDIA’S WESTERN HIMALAYAN REGION: A NAÏVE BAYES CLASSIFIER ANALYSIS

The natural resources-based Microenterprises are the major part of the economy of the western Himalayan region of Uttara hand, India, as the region is predominantly covered with reserved forests. The present study evaluates the performance of Microenterprises and the factors affecting it in the region using the primary data enumerated from 110 microenterprises sampled under four major categories of microenterprises, viz, agro and allied, Animal and allied, handicrafts and handlooms, and miscellaneous. The Naïve Bayes classifier approach has been applied to evaluate the performances (Loss-making, breakeven, profit-making, or high-profit making) of these microenterprises based on their performance determining factors such as ease of raw material availability, level of training received, technological advancement, and the extent of market knowledge, and also on the type of ownership and the employee's number. The Naïve Bayes classification accuracy on the training dataset was 100%, while accuracy on the test dataset ranged from 93% to 100%. The results revealed that agro-based microenterprises have a greater probability (0.67) of making a profit/high profit, while animal product-based microenterprises have a high probability of running into losses. A higher level of Market Knowledge contributes to a high probability (0.89) of making high profits. The higher level of technology and training provides greater chances/probability (0.72, 0.72) of making high profits. Self-help groups (SHGs) have shown a better probability of making profits. The study suggests promoting SHGs in the region, wider dissemination of the market knowledge (marketing strategy), and leveling up the training/technology of the microenterprise.

The LASSIE MPS panel: Predicting externally visible traits in dogs for forensic purposes

Predicting the outward appearance of dogs via their DNA, also known as Canine DNA Phenotyping, is a young, emerging field of research in forensic genetics. The few previous studies published in this respect were restricted to the consecutive analysis of single DNA markers, a process that is time- and sample-consuming and therefore not a viable option for limited forensic specimens. Here, we report on the development and evaluation of a Massively Parallel Sequencing (MPS) based molecular genetic assay, the LASSIE MPS Panel. This panel aims to predict externally visible as well as skeletal traits, which include coat color, coat pattern, coat structure, tail morphology, skull shape, ear shape, eye color and body size from DNA using 44 genetic markers in a single molecular genetic assay. A biostatistical naïve Bayes classification approach was applied to identify the most informative marker combinations for predicting phenotypes. Overall, the predictive performance was characterized by a very high classification success for some of the trait categories, and high to moderate success for others. The performance of the developed predictive framework was further evaluated using blind samples from three randomly selected dog individuals, whose appearance was well predicted.

Perbandingan Metode Klasifikasi Naive Bayes Dan K-Nearest Neighbor Dalam Memprediksi Prestasi Siswa

Student achievement is one of the most important aspects in the field of education, so it is necessary to take action to improve the quality of education in dealing with students who have less achievement so that teachers can identify students who need more guidance. This study aims to predict student achievement by processing students' report cards using the Naïve Bayes classification and k-Nearest Neighbor data mining method approach, and utilizing orange software as a testing tool. The data used is data on grade 1 student scores for the 2019 to 2022 academic year totaling 143 data. The evaluation results using test and score yield CA (Classification Accuracy) values ​​for the Naïve Bayes algorithm 0.909 and the k-NN algorithm 0.848, the F1 value for the Naïve Bayes algorithm 0.858 and the k-NN algorithm 0.820, the Precision value for the Naïve Bayes algorithm 0.862 and the k-NN algorithm 0.821 the Recall value of the Naïve Bayes algorithm is 0.857 and the k-NN algorithm is 0.820. The Naïve Bayes AUC value is 0.909 and the k-NN value is 0.848. Therefore, for the case study, the ROC analysis model that has the best accuracy value is Naïve Bayes because the curve is closest to the coordinate point of 0.1. Based on the analysis of each of these values, the amount of False Negative and False Positive data is close to (Symmetric) and the accuracy value is very high, so accuracy can be used as a reference for algorithm performance. With this research, it is hoped that it will make it easier to determine student potential and can help improve the learning system for students who have difficulty in achieving achievement.

DIABETES MELLITUS ATTRIBUTE CLASSIFICATION USING THE NAIVE BAYES ALGORITHM BASED ON FORWARD SELECTION

Diabetes Mellitus is a chronic condition that frequently results in death. Almost every nation has experienced and contributed to this rise in mortality. Consequently, several researchers are motivated to determine this disease's source and prevent the increase in mortality rates. The research was conducted in the field of informatics in partnership with health professionals to determine the causes of this condition. Many informatics researchers employ machine learning techniques to aid in analyzing existing data. This study suggests feature selection based on forward selection and the naive Bayes classification approach to determine this disease's primary aetiology. The results demonstrate that our proposed strategy can increase the classification accuracy of patients. The performance outcomes improved by 169%. According to this theory, it is also known that the primary cause of this disease is its dependence on body mass index and age. Therefore, additional research must explore these two variables' impact on various other disorders.

Chemical named entity recognition in the texts of scientific publications using the naïve Bayes classifier approach

MotivationApplication of chemical named entity recognition (CNER) algorithms allows retrieval of information from texts about chemical compound identifiers and creates associations with physical–chemical properties and biological activities. Scientific texts represent low-formalized sources of information. Most methods aimed at CNER are based on machine learning approaches, including conditional random fields and deep neural networks. In general, most machine learning approaches require either vector or sparse word representation of texts. Chemical named entities (CNEs) constitute only a small fraction of the whole text, and the datasets used for training are highly imbalanced.Methods and resultsWe propose a new method for extracting CNEs from texts based on the naïve Bayes classifier combined with specially developed filters. In contrast to the earlier developed CNER methods, our approach uses the representation of the data as a set of fragments of text (FoTs) with the subsequent preparati`on of a set of multi-n-grams (sequences from one to n symbols) for each FoT. Our approach may provide the recognition of novel CNEs. For CHEMDNER corpus, the values of the sensitivity (recall) was 0.95, precision was 0.74, specificity was 0.88, and balanced accuracy was 0.92 based on five-fold cross validation. We applied the developed algorithm to the extracted CNEs of potential Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease (Mpro) inhibitors. A set of CNEs corresponding to the chemical substances evaluated in the biochemical assays used for the discovery of Mpro inhibitors was retrieved. Manual analysis of the appropriate texts showed that CNEs of potential SARS-CoV-2 Mpro inhibitors were successfully identified by our method.ConclusionThe obtained results show that the proposed method can be used for filtering out words that are not related to CNEs; therefore, it can be successfully applied to the extraction of CNEs for the purposes of cheminformatics and medicinal chemistry.

Development of binary classification models for assessment of drug-induced liver injury in humans using a large set of FDA-approved drugs

Drug-induced liver injury (DILI) has been identified as one of the major causes for drugs withdrawn from the market, and even termination during the late stages of development. Therefore, it is imperative to evaluate the DILI potential of lead compounds during the research and development process. Although various computational models have been developed to predict DILI, most of which applied the DILI data were extracted from preclinical sources. In this investigation, the in silico prediction models for DILI were constructed based on 1140 FDA-approved drugs by using naïve Bayes classifier approach. The genetic algorithm method was applied for the molecular descriptors selection. Among these established prediction models, the NB-11 model based on eight molecular descriptors combined with ECFP_18 showed the best prediction performance for DILI, which gave 91.7% overall prediction accuracy for the training set, and 68.9% concordance for the external test set. Therefore, the established NB-11 prediction model can be used as a reliable virtual screening tool to predict DILI adverse effect in the early stages of drug design. In addition, some new structural alters for DILI were identified, which could be used for structural optimization in the future drug design by medicinal chemists.

An Innovative Perceptual Pigeon Galvanized Optimization (PPGO) Based Likelihood Naïve Bayes (LNB) Classification Approach for Network Intrusion Detection System

ABSTRACTIntrusion detection and classification have gained significant attention recently due to the increased utilization of networks. For this purpose, there are different types of Network Intrusion Detection System (NIDS) approaches developed in the conventional works, which mainly focus on identifying the intrusions from the datasets with the help of classification techniques. Still, it is limited by the significant problems of inefficiency in handling large dimensional datasets, high computational complexity, false detection, and more time consumption for training the models. To solve these problems, this research intends to develop an innovative clustering-based classification methodology to precisely detect intrusions from the different types of IDS datasets. Here, the most recent and extensively used IDS datasets such as NSL-KDD, CICIDS, and Bot-IoT have been employed for detecting intrusions. Data preprocessing has been performed to normalize the dataset to eliminate irrelevant attributes and organize the features. Then, the data separation is applied by forming the clusters by using an intelligent Anticipated Distance-based Clustering (ADC) incorporated with the Density-Based Spatial clustering of applications with noise (DBScan) algorithm. It helps to find the distance and density measures for grouping the attributes into the clusters, which increases the efficiency of classification. Here, the most suitable optimal parameters are selected using the Perpetual Pigeon Galvanized Optimization (PPGO) technique. The extracted features are used for training and testing the dataset samples.Consequently, the Likelihood Naove Bayes (LNB) classification approach is implemented to accurately predict the classified label as to whether normal or attack. During the evaluation, the performance of the proposed IDS framework is validated and compared using various evaluation metrics. Theresults show that the proposed ADC-DBScan-LNB model outperforms the other techniques with improved performance outcomes.

Missile threat assessment system using Naive Bayes classifier approach

In the modern era, missiles have been used to serve both defence and offense purposes. Numerous research projects have been conducted to improve various aspects of missile systems. One area of main interest regarding the development of missiles in general is automation, particularly surrounding a feature known as guidance system. This feature had been proven to relieve a missile user of several workloads, particularly during aim and launch phase. One key component of a guidance system is known as threat assessment system. Such system is assigned to assess the threat situation that a missile’s engaged in. In this research paper, the developed guidance system uses an approach that is based on the naive Bayes classifier concept. The idea is to establish a system that first describes relationship between a missile and its’ target during mission in the form of a set of combat parameters. These combat parameters are then treated as evidence that is used by the Bayesian algorithm to update an assigned prior belief, yielding an updated posterior belief. The belief comes in a form of probability distribution with each probability representing a predefined threat situation class.

L2b: Lexicon Boosted Bayesian Classification For Popularity Prediction Of Movies With Improved Accuracy Using Twitter Corpus

Social networking plays a vital role in the research phenomenon. It helps in acquiring the most recent news and information consistently with settling on ground-breaking choices and expectations. There are enormous expressions of people opinion in addition to what's circumventing the world. Twitter, is a well known blogging site where individuals place their perspectives and inclinations based on their interest. This paper presents a Lexicon Boosted Bayesian (L2B) classification for popularity prediction of movies based on twitter corpus. The proposed approach predicts the accomplishment of the film utilizing perspective individuals that might actually be accomplished by assessment examination with Naive Bayes classification and Lexicon approach. In addition, it focuses with the representation of information in R language which stands the best in speaking to the investigation in pictorial arrangement and a guide perception that depicts the area of the tweets where it has come from. The result analysis proves the effectiveness of proposed technique in comparison with the existing approaches for predicting the success of the movies.

Development of novel in silico prediction model for drug-induced ototoxicity by using naïve Bayes classifier approach

Some drugs have the potential to cause cellular degeneration of cochlear and/or vestibular system, leading to temporary or permanent hearing loss, innitus, ataxia, dizziness, ear infections, hyperacusis, vertigo, nystagmus and other ear problems. Thus, precise assessment of ototoxicity has become a strong urge task for the toxicologist. In this research, the in silico prediction model of ototoxicity was developed based on 2612 diverse chemicals by using naïve Bayes classifier approach. A set of 7 molecular descriptors considered as important for ototoxicity was selected by genetic algorithm method, and some structural alerts for ototoxicity were identified. The established naïve Bayes prediction model produced 90.2% overall prediction accuracy for the training set and 88.7% for the external test set. We hope the established naïve Bayes prediction model should be employed as precise and convenient computational tool for assessing and screening the chemical-induced ototoxicity in drug development, and these important information of ototoxic chemical structures could provide theoretical guidance for hit and lead optimization in drug design.

Identify the Relevant Pages of Book to be Indexed Using Naive Bayes Classification Method

The back-of-book index is a component that is often found in non-fiction books. The back-of-book index contains important terms and page numbers where the terms appear in the book, it is useful for helping readers find certain term directly, without having to search on every page. Usually, the back-of-book index is compiled by the author or a professional indexer, although currently there are several automatic indexing applications available. To determine a term on a page worth to be indexed requires knowledge and expertise of the author whom better understand the book’s context. Hence, generating a good back-of-book index is a task that requires great effort, knowledge and cost. Therefore, the back-of-book index is very possible containing indexed terms that refer to irrelevant page numbers due to human error or the weakness of the indexing application. This study aims to identify the relevance of the pages to be indexed in the back-of-book index, using Naive Bayes Classification. The testing result shows that the approach with Naive Bayes Classification produces an average precision value of 74.02 percent and 100 percent for the recall value. The average precision value that more than 50 percent indicates that the naive bayes classification approach is capable to identify the relevant and irrelevant pages.

Analysing political events on Twitter

Recently, political events, such as elections, have raised a lot of discussions on social media networks, in particular, Twitter. This brings new opportunities for social scientists to address social science tasks, such as understanding what communities said or identifying whether a community has an influence on another. However, identifying these communities and extracting what they said from social media data are challenging and non-trivial tasks. We aim to make progress towards understanding 'who' (i.e. communities) said 'what' (i.e. discussed topics) and 'when' (i.e. time) during political events on Twitter. While identifying the 'who' can benefit from Twitter user community classification approaches, 'what' they said and 'when' can be effectively addressed on Twitter by extracting their discussed topics using topic modelling approaches that also account for the importance of time on Twitter. To evaluate the quality of these topics, it is necessary to investigate how coherent these topics are to humans. Accordingly, we propose a series of approaches in this thesis. First, we investigate how to effectively evaluate the coherence of the topics generated using a topic modelling approach. The topic coherence metric evaluates the topical coherence by examining the semantic similarity among words in a topic. We argue that the semantic similarity of words in tweets can be effectively captured by using word embeddings trained using a Twitter background dataset. Through a user study, we demonstrate that our proposed word embedding-based topic coherence metric can assess the coherence of topics like humans [1, 2]. In addition, inspired by the precision at k metric, we propose to evaluate the coherence of a topic model (containing many topics) by averaging the top-ranked topics within the topic model [3]. Our proposed metrics can not only evaluate the coherence of topics and topic models, but also can help users to choose the most coherent topics. Second, we aim to extract topics with a high coherence from Twitter data. Such topics can be easily interpreted by humans and they can assist to examine 'what' has been discussed and 'when'. Indeed, we argue that topics can be discussed in different time periods (see [4]) and therefore can be effectively identified and distinguished by considering their time periods. Hence, we propose an effective time-sensitive topic modelling approach by integrating the time dimension of tweets (i.e. 'when') [5]. We show that the time dimension helps to generate topics with a high coherence. Hence, we argue that 'what' has been discussed and 'when' can be effectively addressed by our proposed time-sensitive topic modelling approach. Next, to identify 'who' participated in the topic discussions, we propose approaches to identify the community affiliations of Twitter users, including automatic ground-truth generation approaches and a user community classification approach. We show that the mentioned hashtags and entities in the users' tweets can indicate which community a Twitter user belongs to. Hence, we argue that they can be used to generate the ground-truth data for classifying users into communities. On the other hand, we argue that different communities favour different topic discussions and their community affiliations can be identified by leveraging the discussed topics. Accordingly, we propose a Topic-Based Naive Bayes (TBNB) classification approach to classify Twitter users based on their words and discussed topics [6]. We demonstrate that our TBNB classifier together with the ground-truth generation approaches can effectively identify the community affiliations of Twitter users. Finally, to show the generalisation of our approaches, we apply our approaches to analyse 3.6 million tweets related to US Election 2016 on Twitter [7]. We show that our TBNB approach can effectively identify the 'who', i.e. classify Twitter users into communities. To investigate 'what' these communities have discussed, we apply our time-sensitive topic modelling approach to extract coherent topics. We finally analyse the community-related topics evaluated and selected using our proposed topic coherence metrics. Overall, we contribute to provide effective approaches to assist social scientists towards analysing political events on Twitter. These approaches include topic coherence metrics, a time-sensitive topic modelling approach and approaches for classifying the community affiliations of Twitter users. Together they make progress to study and understand the connections and dynamics among communities on Twitter. Supervisors : Iadh Ounis, Craig Macdonald, Philip Habel The thesis is available at http://theses.gla.ac.uk/41135/

Integrative Analysis Identifies Genetic Variants Associated With Autoimmune Diseases Affecting Putative MicroRNA Binding Sites.

Genome-wide and fine mapping studies have shown that more than 90% of genetic variants associated with autoimmune diseases (AID) are located in non-coding regions of the human genome and especially in regulatory sequences, including microRNAs (miRNA) target sites. MiRNAs are small endogenous noncoding RNAs that modulate gene expression at the post-transcriptional level. Single nucleotide polymorphisms (SNPs) located within the 3′ untranslated region of their target mRNAs (miRSNP) can alter miRNA binding sites. Yet, little is known about their effect on regulation by miRNA and the consequences for AID. Conversely, it is well known that two or more AID may share part of their genetic background. Here, we hypothesized that miRSNPs could be associated with more than one AID. To identify miRSNPs associated with AID, we integrated results from three different prediction tools (Polymirts, miRSNP, and miRSNPscore) using a naïve Bayes classifier approach to identify miRSNPs predicted to affect binding sites of miRNAs. Further, to detect miRSNPs associated with two or more AID, we integrated predictions with summary statistics from 12 AID studies. In addition, to prioritize miRSNPs, miRNAs and AID-associated target genes, we used public expression quantitative trait locus (eQTL) data and mRNA-seq and small RNA-seq data. We identified 34 miRNSPs associated with at least two AID. Furthermore, we found 86 miRNAs predicted to target 18 of the associated gene's mRNAs. Our integrative approach revealed new insights into miRNAs and AID associated target genes. Thus, it helped to prioritize AID noncoding risk SNPs that might be involved in the causal mechanisms, providing valuable information for further functional studies.

An improved model based on genetic algorithm for detecting intrusion in mobile ad hoc network

In any data communication between networks, it is very essential to maintain a high level of security to make sure that the data communication is safe and trusted. There may be chance for intrusions and misuses of information in a network. Intrusion detection systems (IDSs) has become an important component in terms of computer and network security. Detection can be host based or network based. The existing approaches being utilized in intrusion detection are not efficient. This paper presents a genetic algorithm based IDS to detect the malicious nodes in the network. Cloud computing understands users in network nodes in ad hoc or fixed manner to have long term connectivity for service utilization. The proposed approach performs better compared to Gaussian Naive Bayes classifier approach. The objective value of chromosome is calculated through a set of iterations. The experimental results show that the computation time is reduced with high successful detection and network performance is improved to provide reliable transmission. The model is also suitable for users on cloud framework.

Machine Users Detection on Sina Weibo Platform

In recent years, the rapid development of Sina Weibo has made it the representative of many Weibo platforms in China. Sina Weibo has attracted large numbers of users in China because of its fast speed of information dissemination, simple use and many star users. More and more Chinese people get news and share information through Sina Weibo. In addition to the normal users, Sina Weibo also appeared on some machine users, these users are generated in order to create false sentiment, which seriously affected the good order of the Sina Weibo platform. By studying normal users and machine users, this paper extracts eight features, they are the number of followings, the number of followers, the number of Weibos, the number of years using Sina Weibo, Sunshine credit, the number of Weibos you like, the proportion of following others by recommending and the ratio of followings and followers. Naive Bayes classification approach, KNN classification approach and SVC classification approach are used for experiment. The experimental results show that the recall rate of the machine users detection is above 96% and the accuracy rate is above 98%, which validates the validity of the features extracted in this paper.

Predicting drug-induced liver injury in human with Naïve Bayes classifier approach.

Drug-induced liver injury (DILI) is one of the major safety concerns in drug development. Although various toxicological studies assessing DILI risk have been developed, these methods were not sufficient in predicting DILI in humans. Thus, developing new tools and approaches to better predict DILI risk in humans has become an important and urgent task. In this study, we aimed to develop a computational model for assessment of the DILI risk with using a larger scale human dataset and Naïve Bayes classifier. The established Naïve Bayes prediction model was evaluated by 5-fold cross validation and an external test set. For the training set, the overall prediction accuracy of the 5-fold cross validation was 94.0%. The sensitivity, specificity, positive predictive value and negative predictive value were 97.1, 89.2, 93.5 and 95.1%, respectively. The test set with the concordance of 72.6%, sensitivity of 72.5%, specificity of 72.7%, positive predictive value of 80.4%, negative predictive value of 63.2%. Furthermore, some important molecular descriptors related to DILI risk and some toxic/non-toxic fragments were identified. Thus, we hope the prediction model established here would be employed for the assessment of human DILI risk, and the obtained molecular descriptors and substructures should be taken into consideration in the design of new candidate compounds to help medicinal chemists rationally select the chemicals with the best prospects to be effective and safe.