This paper is aimed to identify the key factors on behavioural patterns that affect audiences in making decision to watch films by using the techniques in multiple regressions (MR). The success of a country's film industry is determined by the collaboration between film producers and audiences. Audiences are a major contributor that will determine the success of a film, while producers is the main supplier to provide films to the audiences. Mathematical modelling concept is used in the methodology where two categorical variables, selected after factor analysis, are transformed into dummies. The search for significant factors would involve data cleaning, factor analysis, dummy transformation and four-phase model building. The modelbuilding phases will incorporate the procedures of selecting the best model involving multicollinearity test using the variance-based approach (VIF) below 5. Results revealed that the best model obtained consists of factors on interest, family, friends, the internet and printed medium of posters and brochures which are influential in attracting audiences. This study has therefore extended our knowledge to build bridges between two different fields, that is scientific and non scientific through the concept and methodology of modelling, and consequently, help in advertising and marketing of the film industry by influencing the perceptions of audiences towards cinemas. Contribution/ OriginalityThis study uses new estimation methodology in Social Sciences where significant factors in real-world problems are identified; in this paper, behavioral patterns of film viewers in Malaysia.The mathematical modelling approach with multicollinearity removals and elimination of Humanities and Social Sciences Letters insignificant factors, besides other statistical procedures and tests are incorporated for model"s robustness.

ABSTRAK Jumlah hujan yang tinggi merupakan salah satu penyebab kejadian banjir. Intensiti atau jumlah hujan di satu-satu kawasan pula dipengaruhi oleh fenomena musim monsun. Pantai Timur Semenanjung misalnya mempunyai purata jumlah hujan yang tinggi ketika berlakunya musim monsun timur laut. Keadaan ini menyebabkan berlakunya peningkatan terhadap tahap bahaya banjir ketika musim tersebut. Ini menunjukkan peningkatan tahap bahaya banjir di sesetengah kawasan khususnya di Pantai Timur Semenanjung dipengaruhi oleh kejadian monsun timur laut. Daerah Beaufort merupakan salah satu kawasan hot spot banjir di Sabah. Oleh itu, kajian ini ingin mengenal pasti sama ada kejadian banjir yang berlaku di daerah Beaufort juga dipengaruhi oleh fenomena monsun. Justeru, data hidrologi yang diperoleh dari Jabatan Pengairan dan Saliran (JPS) telah diintegrasikan bersama data pengukuran tikas di kawasan kajian bagi menentukan ciri-ciri banjir sepanjang tempoh 10 tahun (2009-2018). Data dianalisis menggunakan perisian Anaconda Python versi 3.7 melalui aplikasi Pandas. Melalui analisis tersebut, ciri-ciri banjir seperti aras kedalaman, tempoh masa kejadian dan kekerapan banjir diperolehi. Kajian ini mendapati tahap bahaya banjir di daerah Beaufort adalah lebih dominan pada musim monsun timur laut berbanding monsun barat daya. Walhal, purata jumlah hujan didapati lebih tinggi pada musim monsun barat daya. Ini menunjukkan tahap bahaya banjir di daerah Beaufort tidak dipengaruhi sepenuhnya oleh fenomena angin monsun.   ABSTRACT High rainfall is a major cause of floods. The intensity or amount of rainfall in one area is influenced by the monsoon season. The East Coast Peninsular Malaysia, for example, has a high average rainfall during the northeast monsoon season. This has led to an increase in the flood risk during the season. This indicates an increase in the level of flood hazards in some areas, particularly on the East Coast Peninsular, affected by the northeast monsoon event. Beaufort District is one of the flood hot spots in Sabah. Therefore, this study seeks to determine whether flood events occurring in the Beaufort area are also influenced by monsoon phenomena. Therefore, the hydrological data obtained from the Department of Irrigation and Drainage were integrated with the strandline measurement data in the study area to determine the characteristics of the flood over a period of 100 years (2009 to 2018). The data were then analyzed using Anaconda Python version 3.7 software through the Pandas application. Through these analyzes, the characteristics of floods such as depth, duration and frequency of floods are obtained. This study found that the level of flood hazard in the Beaufort area was more dominant during the northeast monsoon than in the southwest monsoon. However, the average rainfall was higher during the southwest monsoon season. This indicates that the level of flood hazard in the Beaufort area is not completely affected by the monsoon phenomenon.