BDI & SCFI prediction (‘22/25W)

In the 25th week, BDI and SCFI predicted values and actual values are compared and tested.

---

1.     Last Week’s forecasts

1-1.        BDI, Actual Vs. Predicted

Date	Last week Forecast	Actual BDI	MAPE(%)	Accuracy(%)
‘22/24W(6/13~17)	2,411.164	2,394.20	0.7%	99.3%

 

1-2.        SCFI, Actual Vs. Predicted

Date	Last week Forecast	Actual SCFI	MAPE(%)	Accuracy(%)
‘22/24W(6/13~17)	4,241.340	4,221.96	0.5%	99.5%

 

 The prediction accuracy of BDI and SCFI was 99.3% and 99.5%, respectively. The predicted value (2,411.164) and actual value (2,394.20) of BDI achieved accuracy and trend, but the predicted value (4,241.340) and actual value (4,221.96) of the SCFI showed a decrease in the actual value, contrary to the expectation that the SCFI would slightly increase. The one-week average of BDI was similar, but it is worth noting that the increase in the second half of the week (Thursday/Friday) is increasing.

  The average BDI MAPE value for the last 12 predictions is 4.7% and the prediction rate is 95.3%, and the SCFI average MAPE value is 0.6% and the prediction rate is 99.4%.

---

2.     BDI forecast

2-1. ARIMA, Time series analysis

 For the BDI forecast, I used weekly data (437 pieces) from January 3, 2014 to Jun 17, 2022 to predict the freight for three weeks in the future.

Date	Forecast	Trend(WoW)
‘22/25W(6/20~24)	2,416.467	↗
‘22/26W(6/27~7/1)	2,517.136	↑
‘22/27W(7/4~8)	2,601.486	↑

 As shown in the table and figure above, BDI (as of 6/17: 2,394.20) declined -1.4% compared to last week (as of 6/10: 2,429.20), narrowing the decline. In particular, the BDI on June 16: 2,462 and June 17: 2,578 rose sharply in the second half of the week. As mentioned in last week's article (week 24), BDI seems to have bottomed out and turned upward.

 The BDI prediction result of the ARIMA model showed that it would rise for 3 weeks in a row.

Ref) BDI & SCFI prediction (‘22/24W) (tistory.com)

BDI & SCFI prediction (‘22/24W)

---

2-2. Logistic regression analysis

 The result of logistic regression analysis was 0.34, which is slightly lower than the baseline of 0.5, but increased significantly from last week's predicted value (0.048). In other words, although a decline is expected, it is necessary to keep an eye on the forecast for next week, as it has largely recovered the extent of the decline.

---

3.     SCFI Outlook

3-1. ARIMA, Time series analysis

 For the SCFI forecast, I used weekly data (635 pieces) from October 16, 2009 to Jun 17, 2022 to predict the future freight for three weeks.

Date	Forecast	Trend(WoW)
‘22/25W(6/20~24)	4,218.104	↘
‘22/26W(6/27~7/1)	4,213.444	↘
‘22/27W(7/4~8)	4,213.711	→

 

 As shown in the table and figure above, the SCFI (as of 6/17: 4,221.96) turned down by -0.27% compared to last week (as of 6/10: 4,233.31). The transition to a decline after ending a four-week rise in recent weeks is very symbolic. In other words, since the extent of the uptrend was very small, it is highly likely that the downtrend in SCFI will continue for the time being due to this week's transition.

---

 

3-2. Logistic regression analysis

 

 The predicted value of the logistic regression analysis model of SCFI was 0.397, which predicted the direction of the SCFI freight rate to decrease next week. As the result of the ARIMA model is the same, the SCFI is highly likely to fall.

---

3-3. SVR model

 

 On the other hand, in the SVM model, unlike the previous models (ARIMA, Logistics), the direction of SCFI was predicted to rise. Although the predicted rise value is not large, it is necessary to look carefully because the direction itself is different. However, since the previous two models judged the direction of the forecast to be down, the author also forecasts the direction of SCFI freight rates to drop next week.

 

Date	Forecast	Trend(WoW)
‘22/25W(6/20~24)	4,238.412	↗
‘22/26W(6/27~7/1)	4,279.054	↗
‘22/27W(7/4~8)	4,318.882	↗

 

Thanks.