Prioritized Mapping in Bali: Multi-View Hierarchical Agglomerative Clustering (MVHAC) Approach

This study aims to determine which sectors and areas should be prioritized in development in BalI. The data used is panel data from 2010 – 2022 for all districts/cities in Bali. The data analysis technique for mapping sectors and regions is Multi-View Hierarchical Agglomerative Clustering (MVHAC). The results of the analysis of Location Quotient and Dyanmic Location Quotient show that the SARBAGITA area (Denpasar, Badung, Gianyar, and Tabanan) has potential in the secondary and tertiary sectors, while other areas are in the primary sector. MVHAC analysis shows that in quadrant 1 (cluster 1), namely Badung and Denpasar, the sectors that can be developed are agriculture, forestry and fisheries; real estate; manufacturing industry; water supply, waste management and waste recycling and mining and excavation, the area of quadrant 3 (cluster 2) of Tabanan and Gianyar is almost similar to Badung and Tabanan, but the potential sector is also transportation and warehousing, while quadrant 4 (cluster 2) is more on the information and communication; real estate; educational services; health services and social activities.


INTRODUCTION
Economic development as a multidimensional process to accelerate economic growth, reducing inequality, and eliminating absolute poverty (Todaro, 1995:21). The substantive freedoms and strengthening the autonomy of individuals able to fully participate in economic activities must be seen as an interrelated process between the factors that produce economic growth nationally and regionally (Sen, 1999;Todaro, 1995:24).
Economic development is a long-term process related to capacities and capabilities among the individual, firms, or industries to achieve economic growth.
Regional development is the key to a country's development because the goals of national development will be achieved if each region can be carried out properly (Pike et al., 2011;Mahendrayasa, 2021). The objective of Gini Ratio (%) regional development is to improve life and equal opportunities (Bachtler & Yuill, 2001;Mustafa, 2002;Andriyansyah et al., 2021). The regional development growth is expected to give impact to all communities in the development area known as inclusive growth (Kim & Daugherty, 2018).
Decentralization or known as regional autonomy is a manifestation of the government system shifting in Indonesia from centralized government system to localized (Maryanov, 2019;Vidyattama et al., 2020).  The economic structure of Bali has a unique characteristic when compared to other provinces, whereas the tourism industry is the leading sector that can encourage economic activity (Picard, 1996). Tourism as the main sector does not always support the uncertain situation and needs support from other sectors (Priatmoko et al., 2021;Subadra & Hughes, 2021). Tourism will not always have a direct effect on inclusive growth in Bali's economic growth, because it is very dependent on regional potential and uncertain situations. Moreover, Bali is unique in terms of culture, customs, and social conditions which also affect the economic activities of the Balinese which are then interesting for discussion. Regional development by optimizing sources used to support economic growth were expected to solve the problem of exclusiveness (Chong, 2020).
The effect of economic growth to income inequality or gini ratio conducted by many researchers. An empirical study of SDGs in Indonesia by Anwar et al (2021) shows if economic growth has an inverse relationship to gini ratio means if the economic growth increases and gini ratio decreases. In the long run, economic growth harm income equality, so the government need to focus on standard policies rather than selective for poor areas (Sari et al., 2021), while Amri & Nazamuddin (2018) show that in the long run there is a negative and significant relationship between the economic growth and income inequality, but in the short run will be positive and insignificant. Mukhlis et al (2018) argue that economic growth does not have any significant effect on income inequality.
Inequality between regions where one region is the leading area and the other is the lagging areas matter to solved.
Handling lagging areas, the government should search for systemic, multisector opportunities, and create a strategy map (Kaplan et al., 2018 (Giannakis & Bruggeman, 2020).
Mapping on priority areas and sectors will reduce an unequal development that will serve as a support for the development of surrounding areas that are still largely underdeveloped (Piętak, 2022). The priority areas should have well-established human resources, nature, infrastructure, and an adequate regional growth center. The adequate access will increase the opportunities of underdeveloped areas and grow to reduce regional disparities.

RESEARCH METHODS
This study aims to satisfy the  (Bickel & Scheffer, 2004;Fernandez & Gomez, 2008;Mirzaei, 2010;Munandar et al., 2018). This is not a new concept, but this concept tried to combine the strengths of Klassen, LQ, DLQ, and HAC methods into a single unit called MVHAC.

Location Quotient (LQ)
This method is usually used to know the direction of regional development based on the sector that is known as potential sector should be developed and it can be used LQ approach. This approach will help to determine which sector should be maintained to support the future development and which sector should be prioritized (Guimarães et al., 2009 (Isserman, 1977). The dynamic location quotient will overcome the static character of the standard of LQ by capturing the changes in sectoral structure over time during the analysis (Goschin, 2021;Pominova et al., 2021). DLQ will be beneficial for identifying the emerging trends in regional specialization and helps estimating future changes in the regional economy (Pribadi & Nurbiyanto, 2021 (Sasirekha & Baby, 2013). This method will be started by calculating the distance between data objects in this research and all of each distance will be seen as a singleton cluster to create a new cluster based on the distance classification (Lukasov, 1978). The equation to calculating the distance is follows: where: Deuc: The euclidean distance between x and y as data object n: The dimension of data number x, y: The data object that will be calculated (first and second data object) 5.

Multi-View Hierarchical
Agglomerative Clustering

(MVHAC)
Muller et al (2012) is proposed the multi view clustering to illustrate the mapping of sectoral development in a particular area. Each region and sector will initially be clustered in the first phase then each group will be clustered again according to its potential sector and GPD growth (Munandar et al., 2018 However, the increase in population has not been followed by an even increase in welfare, this is characterized by a high gini ratio and economic growth that has not been inclusive of inequality. The increase in the population in Bali based on the census results can be seen through the figure below:       and company service are not base sectors but prospective.   Regency.  distance between these regencies is determined by the contribution value of the developing sector in the region and the average growth that occurs in that sector to boost the regional economy.     has the potential to grow in the area.

Klassen Typology Analysis
Then the education services sector; health services and social activities; procurement of electricity and gas; and construction is a very important sector.     Research conducted by Kurniawan & Huda (2020)