China
Africa
Explore chapters and articles related to this topic
The Implementation of the Min-Hashing algorithm in Mahout
Published in Amir Hussain, Mirjana Ivanovic, Electronics, Communications and Networks IV, 2015
Hongya Wang, Xisong Wu, Shan Chang, LihChyun Shu
Figure 2 shows the logical view of the minhash clustering package in Mahout. MinHashDriver is the driver of this clustering algorithm, which initializes parameters of the algorithm and binds the map phase to MinHashMapper and the reduce phase to MinHashReducer. The HashFunction is an interface with a single hash method, which will be overwritten in MinHashMapper. Implementations of different hash functions are defined in HashFactory, which provides the simple factory method for fetching HashFunctions of a requested type, including linear, polynomial, murmur, and murmur3. The core algorithm is implemented in MinHashMapper and MinHashReducer, which will be discussed in details in the next subsection.
Performance analysis and design of competitive business models
Published in International Journal of Production Research, 2018
To find out the potential customer needs, Chiu, Chu, and Chen (2017) integrated the failure modes and effects analysis and important performance analysis with a service–product development matrix. They have verified the proposed approach using the case study of clothing industry. A hybrid modelling approach with self-organising maps for market clustering and an optimisation model for market deployment road mapping within each cluster was proposed by Kimiagari and Montreuil (2017). The proposed approach was applied to the global relief supply venture. The comprehensive review of the previous employed research in the domain of production planning and control functions in SMEs in the era of Industry 4.0. can be seen in Moeuf et al. (2017). Furthermore, the set-up and production planning of hybrid system with manufacturing and remanufacturing was studied for the policy optimization (Polotski, Kenne, and Gharbi 2017). A stochastic model with features of enhancements in energy efficiencies and time-dependent technology cost, vague energy prices, valuable technology life and inconsistent energy consumption was developed by Tan and Yavuz (2015) to present energy-saving technologies as a service. Bullinger and Schweizer (2006) introduced the intelligent competition strategies based on the ‘Digital Factory’ method to overwhelm the cost leadership duality. Spector (2011) integrated the two influential and realistic management theories including theory of constraint and business modelling theory. An exploratory study was conducted by Weerakkody and Irani (2010) to investigate the value and risk of different offshore outsourcing models.
GNSSer: objected-oriented and design pattern-based software for GNSS data parallel processing
Published in Journal of Spatial Science, 2021
Linyang Li, Zhiping Lu, Zhengsheng Chen, Yang Cui, Dashuang Sun, Yupu Wang, Yingcai Kuang, Fangchao Wang
Figure 8 shows the flowchart for GNSS positioning. All the design principles and design patterns mentioned above are used here. In the data loading module, data with standard exchange formats – RINEX, SP3, CLK, ERP, BLQ – are loaded based on the SRP and iterator pattern. This process is highly automatic, for only observation files are mandatory; assisting files are searched, matched in the local directory and downloaded from the International GNSS Service (IGS) file transfer protocol (ftp) automatically. In the pre-processing module, the raw GNSS observations of each epoch are regarded as epoch material based on OCP, and an observation data and product filter is established based on the responsibility chain to filter possible gross errors and check the availability, consistency, accuracy and quality. Moreover, cycle slips can be detected based on the TurboEdit algorithm (Blewitt 1990) and its improvements (Cai et al. 2013). In the error correcting module, four chains are ready to correct the slant distance based on the responsibility chain pattern. The satellite clock and position and the positions of the Sun and Moon are derived based on a singleton pattern. In the parameter module, the adjustment equation constructor and adjustment information processor are available based on the factory method, template method pattern and adapter pattern. Finally, in the results analysis module, epoch results are obtained and checked, and the result files can be stored in the form of SINEX, Excel or HTML, which can satisfy the requirements of different users. To evaluate those results, GNSSer provides plot tools for factors such as the station distribution, satellite visibility, dilution of precision, positioning bias, tropospheric delay, ionospheric delay and ambiguity. All the modules are based on LSP, DIP, ISP and CRP to optimise the architecture of GNSSer.