Dealing with seasonality by narrowing the training set in time series forecasting with k NN

Martínez, Francisco; Frías, María Pilar; Pérez-Godoy, María Dolores; Rivera, Antonio J.

doi:10.1016/j.eswa.2018.03.005

Cited by 69 publications

(25 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…[14] menerapkan pembobotan pada algoritme kNN untuk klasifikasi dengan direct cost sensitive dan distance cost sensitive kNN. Penanganan prediksi data timeseries yang memiliki sifat musiman juga telah dikaji dalam [15]. Penanganannya dengan membagi data training sesuai periode musiman tersebut dan hasil prediksi merupakan agregat dari sejumlah nilai target.…”

Section: Pendahuluanunclassified

Optimization of k value and lag parameter of k-nearest neighbor algorithm on the prediction of hotel occupancy rates

Akbar

Kusumodestoni

2020

Jurnal Teknologi dan Sistem Komputer

View full text Add to dashboard Cite

Hotel occupancy rates are the most important factor in hotel business management. Prediction of the rates for the next few months determines the manager's decision to arrange and provide all the needed facilities. This study performs the optimization of lag parameters and k values of the k-Nearest Neighbor algorithm on hotel occupancy history data. Historical data were arranged in the form of supervised training data, with the number of columns per row according to the lag parameter and the number of prediction targets. The kNN algorithm was applied using 10-fold cross-validation and k-value variations from 1-30. The optimal lag was obtained at intervals of 14-17 and the optimal k at intervals of 5-13 to predict occupancy rates of 1, 3, 6, 9, and 12 months later. The obtained k-value does not follow the rule at the square root of the number of sample data.

show abstract

Section: Pendahuluanunclassified

Optimization of k value and lag parameter of k-nearest neighbor algorithm on the prediction of hotel occupancy rates

Akbar

Kusumodestoni

2020

Jurnal Teknologi dan Sistem Komputer

View full text Add to dashboard Cite

show abstract

“…The historical data of the previous 7 days, 21 days, and 40 days are employed to forecast the system load from January 10 to January 16, 2015. To evaluate the representation learning ability of the deep forest regression for STLF, random forest [56], bagging algorithm [57], gradient boosting algorithm [58], back propagation neural network [59], and k-nearest neighbor algorithm [60] are compared with the proposed deep forest regression in this paper.…”

Section: Examples and Resultsmentioning

confidence: 99%

Deep Forest Regression for Short-Term Load Forecasting of Power Systems

et al. 2020

View full text Add to dashboard Cite

Deep neural networks of deep learning algorithms can be applied into regressions and classifications. While the regression performances and classification performances of the deep neural networks are depending on the hyper-parameters of the deep neural networks. To mitigate the adverse effect of the hyper-parameters for the deep learning algorithms, this paper proposes deep forest regression for the short-term load forecasting of power systems. Deep forest regression includes two procedures, i.e., multi-grained scanning procedure and cascade forest procedure. These two procedures can be effectively trained by two completely random forests and two random forests with the default configuration. Then, the deep forest regression is applied into the short-term load forecasting of power systems. The forecasting performances of deep forest regression are compared with that of numerous intelligent algorithms and conventional regression algorithms under the model with the data of previous 7-day, 21-day, and 40-day. Besides, the forecasting performances of deep forest regression with different parameters are compared. The numerical results show that the deep forest regression with default configured parameters can increase the accuracy of the short-term forecasting and mitigate the influences of the experiences for the configuration of the hyper-parameters of deep learning model. INDEX TERMS Deep forest regression, short-term load forecasting, multi-grained scanning procedure, cascade forest procedure.

show abstract

“…Studies about sales prediction has been done in variate business cases, and variate approach. Generally, prediction techniques can be divided into two major categories: non-algorithmic and algorithmic, or statistical model and computational intelligence [5]. Generally, the non-algorithmic approach uses a simple, fixed formula to predict future conditions.…”

Section: Introductionmentioning

confidence: 99%

Promotion Scenario Based Sales Prediction on E-Retail Groceries Using Data Mining

Muqodas¹,

Kusuma²

2021

IJETAE

View full text Add to dashboard Cite

Mostly in many business cases, sales prediction plays an important role. Production planning is a good example. One aspect which affecting sales forecasting is promotion schedule. Since using promotion is commonly done nowadays, especially in internet business, it is hardly seen a day without promotion in Indonesian e-commerce. Thus, this study discusses about forecasting future sales based on promotion scenario data with main objective is to discover the best machine learning algorithm and model to forecast future sales. Promotion mechanism which employed in this study are price cut, buy 1-get 1, and product bundling. We use 577 data from January 2018 to July 2019 as dataset. We compare kNN, GLM, and SVR as the model predictor to forecast number of transactions in a day. From the experiment k-NN yielded the highest performance ability with squared correlation of 0.938. the worst model predictor for this case is GLM with squared correlation of 0.507. We also determine the best parameter input for each parameter using grid optimization method. We discover 2 is the best k value of kNN and Manhattan distance is the best distance calculation for this case

show abstract

Dealing with seasonality by narrowing the training set in time series forecasting with k NN

Cited by 69 publications

References 48 publications

Optimization of k value and lag parameter of k-nearest neighbor algorithm on the prediction of hotel occupancy rates

Optimization of k value and lag parameter of k-nearest neighbor algorithm on the prediction of hotel occupancy rates

Deep Forest Regression for Short-Term Load Forecasting of Power Systems

Promotion Scenario Based Sales Prediction on E-Retail Groceries Using Data Mining

Contact Info

Product

Resources

About