Paper: A novel method for forecasting time series based on fuzzy logic and visibility graph

Problem

Forecasting time series. 

Other methods' drawback: even though existing methods (exponential smoothing, auto-regression and moving average-MA, ARIMA, maximum entropy method, modified grey model) have a good performance, they are not accurate enough for all the complex situations, because of complicated data structure and an increasing demand for accuracy. 

We improved the accuracy.  

Keywords

forecast construction cost index. fuzzy logic. link prediction 

Related work

1. visibility graph;
2. link prediction: 

Link prediction is used to predict future possible links in the network (E.g., Facebook). Or, it can be used to predict missing links due to incomplete data (E.g., Food-webs – this is related to sampling that Olivia spoke of earlier).

Small-World Phenomenon: 或者叫六度空间，你和任何一个陌生人之间所间隔的人不会超过六个

existing methods: node similarity is defined to estimate the likelihood of the existence of a link between two nodes

3. To handle uncertainties: generalized evidence theory; fuzzy logic. 

fuzzy rules are usually called IF-THEN rules. fuzzy propositions(FP), IF <FP1>, THEN <FP2>

Methodology

historical data -----> visibility graphs --------> link prediction -------> prediction revise based on fuzzy logic. 

phase 1: make predictions by using visibility graph and link prediction. 

i. convert time series into a visibility graph; 

ii. calculate the node similarities. 

iii. make initial predictions

phase 2: revise the predictions based on fuzzy logic. 

??

Results

high flexibility and predictability.