Search results

Organizations in many domains generate a considerable amount of heterogeneous data every day. Such data can be processed to enhance these organizations’ decisions in real time. However, storing and processing large and varied datasets (known as big data) is challenging to do in real time. In machine learning, streaming feature selection has always been considered a superior technique for selecting the relevant subset features from highly dimensional data and thus reducing learning complexity. In the relevant literature, streaming feature selection refers to the features that arrive consecutively over time; despite a lack of exact figure on the number of features, numbers of instances are well-established. Many scholars in the field have proposed streaming-feature-selection algorithms in attempts to find the proper solution to this problem. This paper presents an exhaustive and methodological introduction of these techniques. This study provides a review of the traditional feature-selection algorithms and then scrutinizes the current algorithms that use streaming feature selection to determine their strengths and weaknesses. The survey also sheds light on the ongoing challenges in big-data research.

The amount of features in handwritten digit data is often very large due to the different aspects in personal handwriting, leading to high-dimensional data. Therefore, the employment of a feature selection algorithm becomes crucial for successful classification modeling, because the inclusion of irrelevant or redundant features can mislead the modeling algorithms, resulting in overfitting and decrease in efficiency.

The minimum redundancy and maximum relevance (mRMR) and the recursive feature elimination (RFE) are two frequently used feature selection algorithms. While mRMR is capable of identifying a subset of features that are highly relevant to the targeted classification variable, mRMR still carries the weakness of capturing redundant features along with the algorithm. On the other hand, RFE is flawed by the fact that those features selected by RFE are not ranked by importance, albeit RFE can effectively eliminate the less important features and exclude redundant features.

The hybrid method was exemplified in a binary classification between digits “4” and “9” and between digits “6” and “8” from a multiple features dataset. The result showed that the hybrid mRMR +  support vector machine recursive feature elimination (SVMRFE) is better than both the sole support vector machine (SVM) and mRMR.

In view of the respective strength and deficiency mRMR and RFE, this study combined both these methods and used an SVM as the underlying classifier anticipating the mRMR to make an excellent complement to the SVMRFE.

This paper purposed a multi-facet sentiment analysis system.

Hence, This paper uses multidomain resources to build a sentiment analysis system. The manual lexicon based features that are extracted from the resources are fed into a machine learning classifier to compare their performance afterward. The manual lexicon is replaced with a custom BOW to deal with its time consuming construction. To help the system run faster and make the model interpretable, this will be performed by employing different existing and custom approaches such as term occurrence, information gain, principal component analysis, semantic clustering, and POS tagging filters.

The proposed system featured by lexicon extraction automation and characteristics size optimization proved its efficiency when applied to multidomain and benchmark datasets by reaching 93.59% accuracy which makes it competitive to the state-of-the-art systems.

The construction of a custom BOW. Optimizing features based on existing and custom feature selection and clustering approaches.

The purpose of this paper is to create an automatic interpretation of the results of the method of multiple correspondence analysis (MCA) for categorical variables, so that the nonexpert user can immediately and safely interpret the results, which concern, as the authors know, the categories of variables that strongly interact and determine the trends of the subject under investigation.

This study is a novel theoretical approach to interpreting the results of the MCA method. The classical interpretation of MCA results is based on three indicators: the projection (F) of the category points of the variables in factorial axes, the point contribution to axis creation (CTR) and the correlation (COR) of a point with an axis. The synthetic use of the aforementioned indicators is arduous, particularly for nonexpert users, and frequently results in misinterpretations. The current study has achieved a synthesis of the aforementioned indicators, so that the interpretation of the results is based on a new indicator, as correspondingly on an index, the well-known method principal component analysis (PCA) for continuous variables is based.

Two (2) concepts were proposed in the new theoretical approach. The interpretative axis corresponding to the classical factorial axis and the interpretative plane corresponding to the factorial plane that as it will be seen offer clear and safe interpretative results in MCA.

It is obvious that in the development of the proposed automatic interpretation of the MCA results, the authors do not have in the interpretative axes the actual projections of the points as is the case in the original factorial axes, but this is not of interest to the simple user who is only interested in being able to distinguish the categories of variables that determine the interpretation of the most pronounced trends of the phenomenon being examined.

The results of this research can have positive implications for the dissemination of MCA as a method and its use as an integrated exploratory data analysis approach.

Interpreting the MCA results presents difficulties for the nonexpert user and sometimes lead to misinterpretations. The interpretative difficulty persists in the MCA's other interpretative proposals. The proposed method of interpreting the MCA results clearly and accurately allows for the interpretation of its results and thus contributes to the dissemination of the MCA as an integrated method of categorical data analysis and exploration.

This study investigates how social interaction-oriented content in broadcasters' live speech affects broadcast viewers' purchasing and gift-giving behaviors and how broadcaster popularity moderates social interaction-oriented content's effect on the two different behaviors in live-streaming commerce.

A research model was proposed and empirically tested using a panel data set collected from 537 live streams via Douyin (the Chinese version of TikTok), one of the most popular live broadcast platforms in China. A fixed-effects negative binomial regression model was used to examine the proposed research model.

This study's results show that social interaction-oriented content in broadcasters' live speech has an inverted U-shaped relationship with broadcast viewers' purchasing behavior and shares a positive linear relationship with viewers' gift-giving behavior. Furthermore, broadcaster popularity significantly moderates the effect of social interaction-oriented content on viewers' purchasing and gift-giving behaviors.

This research enriches the literature on live-streaming commerce by investigating how social interaction-oriented content in broadcasters' live speech affects broadcast viewers' product-purchasing and gift-giving behaviors from the perspective of broadcast viewers' attention. Moreover, this study provides some practical guidelines for developing live speech content in the live-streaming commerce context.

A new research domain known as the Quantified Self has recently emerged and is described as gaining self-knowledge through using wearable technology to acquire information on self-monitoring activities and physical health related problems. However, very little is known about the impact of time window models on discovering self-quantified patterns that can yield new self-knowledge insights. This paper aims to discover the self-quantified patterns using multi-time window models.

This paper proposes a multi-time window analytical workflow developed to support the streaming k-means clustering algorithm, based on an online/offline approach that combines both sliding and damped time window models. An intervention experiment with 15 participants is used to gather Fitbit data logs and implement the proposed analytical workflow.

The clustering results reveal the impact of a time window model has on exploring the evolution of micro-clusters and the labelling of macro-clusters to accurately explain regular and irregular individual physical behaviour.

The preliminary results demonstrate the impact they have on finding meaningful patterns.

Indian Premier League (IPL) is one of the more popular cricket world tournaments, and its financial is increasing each season, its viewership has increased markedly and the betting market for IPL is growing significantly every year. With cricket being a very dynamic game, bettors and bookies are incentivised to bet on the match results because it is a game that changes ball-by-ball. This paper investigates machine learning technology to deal with the problem of predicting cricket match results based on historical match data of the IPL. Influential features of the dataset have been identified using filter-based methods including Correlation-based Feature Selection, Information Gain (IG), ReliefF and Wrapper. More importantly, machine learning techniques including Naïve Bayes, Random Forest, K-Nearest Neighbour (KNN) and Model Trees (classification via regression) have been adopted to generate predictive models from distinctive feature sets derived by the filter-based methods. Two featured subsets were formulated, one based on home team advantage and other based on Toss decision. Selected machine learning techniques were applied on both feature sets to determine a predictive model. Experimental tests show that tree-based models particularly Random Forest performed better in terms of accuracy, precision and recall metrics when compared to probabilistic and statistical models. However, on the Toss featured subset, none of the considered machine learning algorithms performed well in producing accurate predictive models.

The authors enrich and extend the existing institutional anomie theory (IAT) in the hope of sharpening the understanding of the joint effects of selected cultural values and social institutional changes on women's pre-entrant entrepreneurial attempts. The authors theorize that women are culturally discouraged to pursue pre-entrant entrepreneurial attempts or wealth accumulation in a specific culture. This discouragement creates an anomic strain that motivates women to deviate from cultural prescriptions by engaging in pre-entrant entrepreneurial attempts at a faster speed. Building on this premise, the authors hypothesize that changes in social institutions facilitate the means of achievement for women due to the potential opportunities inherent in such institutional changes.

Using a randomly selected sample of 1,431 registered active individual users with a minimum of 10,000 followers on a leading entertainment live-streaming platform in the People's Republic of China, the authors examined a unique mix of cultural and institutional changes and their effects on the speed of women's engagement in live-streaming platform activity.

The authors find support for the impact of the interaction between changes in social institution conditions and cultural values. Unexpectedly, the authors also find a negative impact of cultural values on women's speed of engaging in pre-entrant entrepreneurial attempts.

The authors add institutional change to the IAT framework and provide a novel account for the variation in the pre-entrant entrepreneurial attempts by women on the platform.

This study embraces the call for exploring the determinants of continuous intention in TikTok. Taking the perspective of social influence, this study not only tries to explore the contextual sources of two types of social influence but also aims to unveil the influence mechanism of how social influence affects TikTok viewers’ continuous intention.

This study empirically analyzes how TikToker attractiveness, co-viewer participation, platform reputation and content appeal affect informative and normative social influence and then lead to the continuous intention of TikTok. Based on 547 valid survey data, this study adopts a mixed analytical approach for data analysis by integrating structural equation modeling (SEM) and fuzzy-set qualitative comparative analysis (fsQCA).

SEM results unveil that content appeal is the most critical antecedent of informational social influence, while the TikToker attractiveness and platform reputation have no effect on it. Differently, all four external sources positively lead to normative social influence. Among them, content appeal and co-viewer participation influence the most. The influences of both two types of social influence on continuous intention are demonstrated. FsQCA results reveal seven alternative configurations that are sufficient for influencing continuance intention and further complement and reinforce the SEM findings.

Addressing the critical contextual elements of TikTok, this study explores and confirms the sources which may engender social influence. The authors also demonstrate the critical role of social influence in affecting TikTok viewers’ continuous intentions by the hybrid analytical approach, which contributes to existing academic literature and practitioners.

The purpose of the study is to dive into various binge-watching habits of Netflix users amidst the COVID-19 pandemic. Consumers find themselves amidst the COVID-19 lockdown with more free time to indulge in these viewing habits. This study investigates motivational factors, amount of media consumption and negative attributes associated with binge-watching on Netflix during the COVID-19 outbreak.

This study has employed an exploratory research design and obtained primary data via an online survey using a semistructured questionnaire. Convenience sampling has been used to choose a sample (n = 105) of Netflix binge-watchers during the COVID-19 pandemic. Both sample selection and survey administration have been done through social media messaging services owing to the COVID-19 lockdown measures.

The results indicate that most of the respondents use smartphones for binge-watching on Netflix. Moreover, they have expressed that a wide range of shows available on Netflix incline them to engage in marathon viewing. However, the respondents spend just over 70 h per month binge-watching on Netflix. Finally, the majority of respondents have flagged “one more episode” syndrome as the most challenging aspect of being marathon viewers on Netflix during the COVID-19 pandemic.

This is one of the few papers to exclusively focus on the impacts of binge-watching on Netflix during the COVID-19 pandemic. The findings will originate the value with novelty and important implications to the Netflix consumers, telecom service providers and payment gateways.