Browsing by Author "Vainio, Tommi"

Filter results by typing the first few letters
Now showing 1 - 9 of 9
  • Maakaasun kulutuksen aikasarjojen analysointi
    (1993) Vainio, Tommi
     Helsinki University of Technology | Master's thesis
  • Kaksifaasisen polypropeenipohjaisen termoplastisen elastomeerin kompatibilisointi
    (1995) Jukarainen, Harri
     Helsinki University of Technology | Master's thesis
  • Due diligence: a problematic view
    (2008) Vainio, Tommi
     School of Business | Master's thesis
  • Integration of Machine Vision to Fiber Optic Cable Manufacturing Lines
    (2000) Leinvuo, Joni
     Helsinki University of Technology | Master's thesis
  • Mixing and Processing of Multiphase Polymer Blends
    (1993) Vainio, Tommi
     Helsinki University of Technology | Licentiate thesis
  • Modeling space elasticity of demand to support retail replenishment planning
    (2018-06-11) Vainio, Tommi
     Perustieteiden korkeakoulu | Bachelor's thesis
  • On Hierarchical Campaign Forecasting
    (2021-03-16) Vainio, Tommi
     Perustieteiden korkeakoulu | Master's thesis
    As retail world is growing more and more competitive, retailers are forced to optimize their operations as well as possible. Accurate demand forecasting is pivotal in optimizing and automizing the replenishment of a retail chain. However, many different factors, such as promotional activities, induce drastic variation in the demand, making forecasting difficult if no appropriate methods or sufficient amount of data is available for the retailer. This thesis specifies different designs of hierarchical, multilevel regression models for demand and campaign effect forecasting utilizing a natural product hierarchy within a retail chain. The performance of the models is evaluated with a dataset obtained from a large European retailer. The model evaluation is three-fold, consisting of assessing the model estimation times, forecast accuracy, and the number of excess errors produced by each model. A particular focus is on cases where there is little data of promotional activities on the lowest hierarchy level, which is a case where conventional models struggle the most. Both frequentist and Bayesian approach is experimented. For the models specified in Bayesian fashion, both Maximum a Posteriori estimation of the parameters and full posterior distribution estimation with Variational Inference methods are implemented, and the results are compared. The thesis aims to find out whether the specified hierarchical models outperform regular, single level Generalized Linear Models specified similarly and to understand the advantages and disadvantages of the differently specified hierarchical models. Also, differences in the results of different estimation algorithms and regression types applicable to the problem are of interest. It is concluded that the parameter estimation of most of the hierarchical models evaluated here is computationally too heavy to be used in operational demand forecasting of large retail chains, while producing little advantages compared to the single-level models. However, Empirical Bayes models seem very promising, as they outperform both the benchmark and the other hierarchical models in almost every evaluated attribute. In addition, a significant finding is that no clear advantages are obtained by using non-normal regression models, in comparison to linear regression where the objective variable is assumed to be normally distributed.
  • The solution buying behavior of Private and Government-owned companies in China
    (2012) Qiao, Liang
     School of Science | Master's thesis
    This study examines the Chinese customer's buying behavior when they do the complex solution buying. Their buying behavior is studied from buying center, buying process, decision making and buying criteria four aspects. The buying behavior difference between Government-owned company and private company are also studied during the research. The objective is to find the common pattern and difference of Chinese customer's buying behavior. Thus, the theory model from literature review can be verified, and practical recommendation and suggestions can be given to case company. The objective is achieved by mainly answer the research questions: 1. What's the adopted theory model for analyzing organizational buying behavior of this case?. 2. How do Chinese customers buy complex solutions?. 3. The difference buying behaviors of government-owned and private companies. 4. What are recommendations and suggestions for case company to do solution selling?. The research methodology is mainly intensive literature reviewing and descriptive multi-case study. Literature review is based on the three fundamental theory models from Robinson, Faris and Wind (1967), Webster and Wind (1972), Sheth (1973). Then the buying behavior theory evolution and latest development are further invested. Practical case study chooses four customer companies which are Government-owned or private companies. The findings from interview, observation and documents are synthesis cross cases. At last, the recommendation and suggestion are given. This study refines the four key aspects of customer's buying behavior. They are then used as the theory model for practical case study. The fruitful findings from practical case study prove that those four aspects model are the good method to have insight of customer's buying behavior. If sales want to promote their product and close the deal, they need to understand their customers in those four aspects. Some unique characteristics of Chinese government and private companies in buying behavior are identified and discussed. The future research direction are given which are the buying difference of government-owned and private companies in emerging countries and culture influence on buying behavior.
  • Tuotesuunnittelun muutoksenhallinta
    (2013-11-18) Parvela, Toni
     Sähkötekniikan korkeakoulu | Master's thesis