Title: Estimating Tea Stock Values Using Cluster Analysis

Year of Publication: Nov - 2014
Page Numbers: 1-11
Authors: Amitha Caldera, Sajitha N. Kaluarachchi, Dilini T. R. Serasinghe
Conference Name: The International Conference on Data Mining, Internet Computing, and Big Data (BigData2014)
- Malaysia


Asia Siyaka is a leading tea broker which has about 15% market share and annually sells tea with the worth of about 18 billion rupees. Tea brokers auction tea on behalf of the factory owners. Many factories buy tea from non-factory holding growers but it takes time for green leaves to become cash at the auction. Hence factories borrow money from the brokers on behalf of future stock to be auctioned to ensure seamless cash flow. The brokers’ main challenge is to estimate the stock which will be sold in the future auctions before granting advances. The current system uses previous month’s realized auction average prices of each factory to estimate the each factory’s stock. This method is incorrect due to the variation of the grade mix of factory wise production and the price variation of tea grades with time without any pattern. Therefore requirement exists to discover an accurate method to estimate the tea stock. This paper used initial descriptive statistics which helped to understand the data and the current system. The paper also used more explorative approach because of the complexity of the problem and unclearness about the relationship of the data. Therefore cluster analysis, which is an unsupervised learning technique, was selected. The analysis shows that comparison should be done with the most recent auction data and also shows that factory, grade and package weight are the only visible attributes which can contribute to future stock value calculations.